6.31 released

Fri Aug 14, 2015 4:50 pm

What's new in 6.31 (2015-Aug-14 15:42):

*) check-for-update - added ability to select versions channel to check
(bugfix, current, RC or development)
*) demo mode of Cloud Hosted Router (CHR) added
*) chr - added x86_64 image for use in virtual environments
*) chr - added support for VMware SCSI virtual disks
*) chr - added support for VMware vmxnet3 network card
*) chr - added support for HyperV SCSI disks
*) chr - added support for HyperV Ethernet interfaces
*) chr - added support for virtio disks
*) fixed occasional interface resetting on CRS switches
*) fixed ethernet stopping on RB NetMetal / SXTG-5HPacD 10Mbit and 100Mbit links
*) ipsec - fixed crash in when gcm encryption was used
*) ipsec - allow to set peer address as "::/0"
*) ipsec - fixed empty sa-src address on acquire in tun mode
*) ipsec - show proposal info in export ipsec section
*) ipsec - preserve port wildcard when generating policy without port override
*) ipsec - fixed replay window, was accidentally disabled since version 6.30;
*) certificate manager - fixed memory leak
*) ssh - allow host key import/export
*) ssh - use 2048bit RSA host key when strong-crypto enabled
*) ssh - support RSA keys for user authentication
*) conntrack - fixed problem with manual connection removal
*) conntrack - added tcp-max-retrans-timeout and tcp-unacked-timeout
*) wireless - implemented l2mtu update if wireless-cm2 is enabled
*) wireless - improved WMM-PowerSave support in wireless-cm2 package
*) mpls - better multicore support for VPLS ingress/egress
*) ovpn - better multicore support for interface initialization/authentication/creation.
*) mesh - performance improvement
*) pptp & l2tp - fixed problem where android client could not connect if both dns names were not provided (was broken since v6.30)
*) user-manager - fixed username was not shown in /tool user-manager user
*) user-manager - fixed zoom for user-manager homepage when mobile devices used
*) winbox - restrict change dynamic interface fields
*) winbox - also hide passphrase in CAPsMAN with "Hide Password"
*) winbox - restrict reversed ranges in dst-port under firewall
*) quickset - fixed HomeAP mode
*) lcd - added LCD package for all architectures (for serial port LCD modules)
*) lcd - fixed crash (and 100% cpu usage) when interface gets removed from "stats-all" screen
*) tool fetch - fixed incomplete ftp download
*) tool fetch - don't trim [t]ftp leading slashes
*) proxy - adjust time according to time-zone settings in proxy cache contents.
*) bridge fastpath - fixed updating bridge FDB on receive (could cause TX traffic flooding on all bridge ports)
*) bonding fastpath - fixed possible crash when bonding master was also a bridge port
*) route - fixed crash on removing route that was aggregated
*) romon - fixed crash on SACKed tx segments
*) lte - improved modem identification to better support multiple identical modems
*) snmp - fixed system scripts table
*) traffic flow - fixed dynamic input/output interface reporting
*) ipv6 dhcp-relay - fixed problem loading configuration

known issue:
*) Dynamic DNS servers can disappear when "allow-remote-requests" are not enabled

Bigfoot · Fri Aug 14, 2015 6:04 pm

Thx for fix this problem on 6.31
*) proxy - adjust time according to time-zone settings in proxy cache contents.
I have tested it and is working correctly.

Bigfoot

marrold · Fri Aug 14, 2015 7:17 pm

Thanks for the update!

*) conntrack - fixed problem with manual connection removal

Perfect timing, I need to debug an issue this weekend and this was preventing me from doing so.

*) wireless - improved WMM-PowerSave support in wireless-cm2 package

Does this need to be enabled in any way?

Have bridges been fixed on the CHR?

dgonzalezh · Fri Aug 14, 2015 7:21 pm

@normis, Thanks for the announcement, I just opened the CCR1009 I got from you after my presentation at the MUM and I'll test 6.31 ASAP.

I'll leave my CRS with the reliable 6.28 as it's serving some clients after I test with CCR, I'll switch and let you know.

Finally it's great to know that the CHR is now available, I'll test.and share my thoughts on it.

Thumbs up!. Thanks Mikrotik

finalcutroot · Fri Aug 14, 2015 8:08 pm

Thank You ,,

i wonder if there is any upcoming updates regarding to hotspot ?? in security specially like preventing Mac cloning

Fri Aug 14, 2015 8:09 pm

Just upgraded one of my RBs, and I can already point a first (minor) issue:

In Winbox, if you then check for updates again against the "current" channel, you always get "Download" and "Download & Install" buttons, but never "Check for updates". This would be OK if there was 6.32, but right now, 6.31 is the latest current, so effectively, Winbox is telling me to download and install the release that I already have installed... which it shouldn't.

mmabob · Fri Aug 14, 2015 8:44 pm

After updating to 6.31 on a pair of RB750 routers, I now have Issues with the ovpn connection between them.

The branch office router that connects to the main router will connect enough that I can send ping traffic over it, but as soon as I try to use the connection for useful data such as connecting to a computer on the other side of the connection, I see in the log "ovpn-out1: terminating... - TLS failed"

I don't know if it is the slightly larger amount of data being transferred that is relating to the issue or what the problem is. It worked well on version 6.27 which we were running before this.

Is anyone else seeing this?

I appreciate any of your help. I currently have a branch office without a useful connection.

patrick7 · Fri Aug 14, 2015 9:37 pm

If I set Update channel to Bugfix and execute /system package update download, it will download the latest Bugfix, instead of nothing.

vortex · Fri Aug 14, 2015 9:50 pm

On webfig you can see that if you set the channel to bugfix, the proposed update is a downgrade version.

This also seemed wrong to me, if you are already on 6.31, the bug fix channel is 6.31.x

In any case, I upgraded my RB2011 from webfig and it seems fine so far. My TV got the right time for now.

patrick7 · Fri Aug 14, 2015 9:52 pm

Yes. It should show a message like "no update available", instead of showing / downloading a previous version.

Except of this, I didn't find any other problems.

105547111 · Fri Aug 14, 2015 9:53 pm

2hrs 22 minutes uptime - stable looks real good

idorotic · Sat Aug 15, 2015 12:21 am

Not working on my 450g, it restarts every few minutes.
On second router 2011 works fine.

elgrandiegote · Sat Aug 15, 2015 5:53 am

OpenVPN Error:
ovpn,debug,error,,,,,,,,,l2tp,info,,debug,,,critical,,,,,,,,,,,,,warning: duplicate packet, dropping

any idea about this?
Version 6.30.2 Works fine

barkas · Sat Aug 15, 2015 11:13 am

Not working on my 450g, it restarts every few minutes.
On second router 2011 works fine.

Same here, rebooting on 450g, stable on 850g.

This is not acceptable, even at home I could not run a Mikrotik setup without redundancy preferably with different models.

marrold · Sat Aug 15, 2015 12:09 pm

This is not acceptable, even at home I could not run a Mikrotik setup without redundancy preferably with different models.

Err, this is why they released a 'stable, bug-fix branch' and then the 'updates' branch.

G2Dolphin · Sat Aug 15, 2015 1:06 pm

Will these firewall fasttrack dummy rules be allowed to be switched off?

Sat Aug 15, 2015 3:14 pm

If I set Update channel to Bugfix and execute /system package update download, it will download the latest Bugfix, instead of nothing.

You are missing the point here. As was explained before, you should consider "Bugfix only" branch as stable. "Current" is where new features are added first, and (since new features often break existing features in some way) is not initially considered stable.

It is rather good Mikrotik now provides an easy way to switch between bugfix (stable) and current. Whether it should or should not allow redownloading/reinstalling the version that is currently installed on your device, is arguable. I'd rather leave it as it is now.

JanezFord · Sat Aug 15, 2015 3:33 pm

Upgraded two of my RB450G and both of them run for 15 hours now ... no reboots, no issuses.... both of them run ipsec, l2tp, ipip, ospf, capsmanv2, about 15 queues and about 60 firewall rules ... Also upgraded 850, 2011, 751, 951, 800, 711, 912 and crs125 devices without problems to 6.31.

JF.

blajah · Sat Aug 15, 2015 3:57 pm

RB951G-2ND and RB750G had a NTP client running before the upgrade. After the upgrade, in profiler NTP process was always on 0.5% ( i have waited 10-15 min). After disabling then enabling NTP client, process appears then disappears from profiler after 10-15 seconds (as expected).

patrick7 · Sat Aug 15, 2015 4:35 pm

If I set Update channel to Bugfix and execute /system package update download, it will download the latest Bugfix, instead of nothing.
You are missing the point here. As was explained before, you should consider "Bugfix only" branch as stable. "Current" is where new features are added first, and (since new features often break existing features in some way) is not initially considered stable.

It is rather good Mikrotik now provides an easy way to switch between bugfix (stable) and current. Whether it should or should not allow redownloading/reinstalling the version that is currently installed on your device, is arguable. I'd rather leave it as it is now.

I'm missing nothing. If I'm on 6.31 and set it to "bugfix only", it should NOT download 6.30.2 (we are in the upgrade menu, not downgrad). It should download bugfix updates of the CURRENT INSTALLED release, that would be 6.31.1. So since this version is not available, it shouldn't download anything.

G2Dolphin · Sat Aug 15, 2015 5:00 pm

I'm missing nothing. If I'm on 6.31 and set it to "bugfix only", it should NOT download 6.30.2 (we are in the upgrade menu, not downgrad). It should download bugfix updates of the CURRENT INSTALLED release, that would be 6.31.1. So since this version is not available, it shouldn't download anything.

If you're on 6.31 — you are on a "latest" release line, and I don't see something strange in that situation, that when you "swtich" the lines — you're getting the last firmware of another line.

dadoremix · Sat Aug 15, 2015 5:17 pm

i try upgrade my RB493aH .. run 5 min and blocked, cant acess via mac or ip..
power off, power on.. again 3min, again 20 min.. WTF with 6.31.. with problems i return to 6.29.1, i have lucky that these mtik was on my house, so i can unplug power, but we have some network in the city, not good option.

patrick7 · Sat Aug 15, 2015 5:45 pm

We are talking about the upgrade menu. Not "Release Switch menu".

If you are on 6.30, "Bugfix" should bring Updates to 6.30.1, 6.30.2, 6.30.3
If you are on 6.31, "Bugfix" should bring Updates to 6.31.1, 6.31.2, 6.31.3

Not downgrades. Even if you download the "downgrade" and reboot, it will not be installed as it is an older version than the installed one. You will have to run /system package downgrade additionally. So in my opinion this feature is not here to switch between releases, but for changing update channel.

teasertime · Sat Aug 15, 2015 5:49 pm

Is this different from the existing x86 build for PC's found on the download page?
*) chr - added x86_64 image for use in virtual environments

Is this the "synthetic" network interface in hyper-v?
*) chr - added support for HyperV Ethernet interfaces

Sat Aug 15, 2015 6:30 pm

I'm missing nothing. If I'm on 6.31 and set it to "bugfix only", it should NOT download 6.30.2 (we are in the upgrade menu, not downgrad).

The menu is called "/system package update" in terminal, and "System -> Packages" in Winbox. The command is called "check-for-updates". Do you feel the difference between "Update" and "Upgrade"?

npero · Sat Aug 15, 2015 7:44 pm

Be very careful with this release now lost one wireless link with Groove52 not come back after update.
Also on 750UP try it to disable wireless package also lost device but they reboot after couple minutes by him self with log message kernel failure but wireless not disabled.
One 750GL every two minute in log say interface,warning ether5-slave-local excessive broadcasts/multicasts, probably a loop.

I update lot of device in last couple years this is maybe second time to lost device after update, that Groove is just simple wireless bridge but not have access in next 7 days

.

patrick7 · Sat Aug 15, 2015 8:05 pm

Neither Upgrade nor Update is Downgrade.

sanitycheck · Sat Aug 15, 2015 9:38 pm

I see something unusual with SSH after the upgrade to 6.31.

I connect to a remote router by SSH using Putty in Windows. A special user is configured for remote SSH access using no password but with a DSA key.

After the upgrade Putty prompted me with the standard warning I get (and expect) when I connect by SSH to a router for the first time. The warning is that because you have never connected before you may not trust the connection, and it shows the fingerprint. After connecting once I never get the warning again for that router.

But after the upgrade to 6.31 I received the warning again, as if I had never connected to that router. Of course I was just in there a minute earlier to trigger the router upgrade. I accepted the connection and it seems to be fine now, but something changed with SSH and 6.31. Not sure if this is a problem or not.

patrick7 · Sat Aug 15, 2015 11:20 pm

Just saw memory usage got bigger and bigger every minute - memory leak again? (had 87MB free RAM) (CCR1009-8G-1S-1S+). Anyone else?

napismizpravu · Sat Aug 15, 2015 11:46 pm

firewall problem not block NAT connections IP time limit expired lifetime in src-address-list...IP removed list in time connections no block, RouterOS 6.31 (uptime 1 day)

v.6.31rc16 autoblock IP works OK function

ip firewall nat print
chain=dstnat action=dst-nat to-addresses=192.168.1.1 to-ports=2345 protocol=tcp src-address-list=WWW in-interface=ether3 dst-port=2345 log=yes log-prefix=""

address list name WWW time limit (D)
ether3 WAN

RB433UAH

idorotic · Sun Aug 16, 2015 12:29 am

Could not add VirtualAP via winbox on my RB751G-2HnD, "L2 MTU" field is empty and grey (read only) so i get error message: "Couldn't add New Interface - l2mtu wrong (6)". It is only possible to add interface via terminal. I tried with winbox 2.2.18 and 3.0rc12.
On my second router RB2011 I can normally add VirtualAP via winbox.

docmarius · Sun Aug 16, 2015 12:41 am

CA certificate import still doesn't work (was ok in 6.29.1).

Sun Aug 16, 2015 12:54 am

Got bootloop after update to 6.31 from 6.30.1 on rb532a. Will need to use a serial cable to check what to do...

moep · Sun Aug 16, 2015 1:10 am

working fine on several devices (CCR, CRS, Groove, 951G, 751G, hap-lite)

one common issue
/system shutdown via Winbox and/or console, ssh, telnet is not working anymore
The device does make a reboot instead.
This issue is found on the platforms mips-be and smips (tilera not tested but likely also affected).
Edit Issue is also present on ppc platform (RB1200 and RB1100AH tested).
The issue is gone after downgrading to 6.30.2. Edit end

I hope this is a bug. Or is shutdown not supported anymore?

zervan · Sun Aug 16, 2015 9:10 am

Just saw memory usage got bigger and bigger every minute - memory leak again? (had 87MB free RAM) (CCR1009-8G-1S-1S+). Anyone else?

No problem here on the same CCR model...

docmarius · Sun Aug 16, 2015 9:14 am

Got reboots every 5 minutes or so on my OmniTIK UPA-5HnD (no auto-spout.rif even if it is enabled in watchdog, free memory is ok, cpu at most at 3-4%).
6.30.2 seems stable.

defekt · Sun Aug 16, 2015 11:48 am

I have issue with switch port flapping with 6.31 on "model: 750", after downgrade to 6.30.2 all is ok

npero · Sun Aug 16, 2015 12:17 pm

750UP upgraded yesterday, today reset by watchdog, never before have that much problem with new release with simple configuration in this case 750UP is just poe switch.
Never in last two years have that much problem with devices with simple configuration.
Yes I know is my mistake to upgrade without testing but not expected problem in devices with basic configuration.
For now not have problem with BaseBox, SXTLite, 411, one Groove52 is lost after upgrade others is working.

bajodel · Sun Aug 16, 2015 1:21 pm

.. ..One 750GL every two minute in log say interface,warning ether5-slave-local excessive broadcasts/multicasts, probably a loop..

The same for me on an CRS 125 on the master-port ..every two minute ..excessive broadcasts/multicasts, probably a loop.
Back to 6.30.2 problem disappear

bigears · Sun Aug 16, 2015 2:08 pm

upgraded to 6.31 on CRS125. IP phones cannot connect to gateway after CRS125 reboot. Had to manually restart each phone. Not acceptable so reverted to 6.30.2. All fine again.

eworm · Sun Aug 16, 2015 2:38 pm

I see something unusual with SSH after the upgrade to 6.31.

I connect to a remote router by SSH using Putty in Windows. A special user is configured for remote SSH access using no password but with a DSA key.

After the upgrade Putty prompted me with the standard warning I get (and expect) when I connect by SSH to a router for the first time. The warning is that because you have never connected before you may not trust the connection, and it shows the fingerprint. After connecting once I never get the warning again for that router.

But after the upgrade to 6.31 I received the warning again, as if I had never connected to that router. Of course I was just in there a minute earlier to trigger the router upgrade. I accepted the connection and it seems to be fine now, but something changed with SSH and 6.31. Not sure if this is a problem or not.

RouterOS now supports RSA keys. Looks like Putty defaults to RSA when available, and it prefers that over the already known DSA key. So everything is just fine.

Either force Putty to use the DSA host key or accept the new RSA key. I would prefer second option.

eworm · Sun Aug 16, 2015 2:43 pm

I opened a thread about high CPU load with wireless-cm2 on version 6.31rc16. This is still valid, though I know the cause and workaround: wireless-cm2 hangs when older CAPs (version <= 6.30) try to update from CAPsMAN. Updating the CAPs first (manually) and the CAPsMAN after works around this.

npero · Sun Aug 16, 2015 3:07 pm

Big WARRNING DO NOT upgrade 750UP I have problem on network after couple hours devices just crash some reset by watchdog some not.
I kill my network with this upgrade.

Sun Aug 16, 2015 7:04 pm

firewall rule CONNECTION RATE option not working

http://wiki.mikrotik.com/wiki/Manual:Connection_Rate

freemannnn · Sun Aug 16, 2015 7:07 pm

please fix connection-rate problem in next bug release!

http://forum.mikrotik.com/viewtopic.php ... 12#p495212

athlonxp78 · Sun Aug 16, 2015 8:47 pm

Big WARRNING DO NOT upgrade 750UP I have problem on network after couple hours devices just crash some reset by watchdog some not.
I kill my network with this upgrade.

+1 The same here in ALL our RB750UP (before 6.30.2), no supout file was posible

elgrandiegote · Sun Aug 16, 2015 9:30 pm

After updating to 6.31 on a pair of RB750 routers, I now have Issues with the ovpn connection between them.

The branch office router that connects to the main router will connect enough that I can send ping traffic over it, but as soon as I try to use the connection for useful data such as connecting to a computer on the other side of the connection, I see in the log "ovpn-out1: terminating... - TLS failed"

I don't know if it is the slightly larger amount of data being transferred that is relating to the issue or what the problem is. It worked well on version 6.27 which we were running before this.

Is anyone else seeing this?

I appreciate any of your help. I currently have a branch office without a useful connection.

Problems with ovpn in this version!!!
on 6.30.2 work fine.

lyovav22 · Sun Aug 16, 2015 9:34 pm

Hi Incoming traffic and CPU up to 100%

MAYAKNET · Sun Aug 16, 2015 9:39 pm

I do not speak English. I will try to express the idea. In release 6.31 instead of 32 cores in a server - it showed only one core and load at rush hour on 100%.

Mon Aug 17, 2015 7:29 am

Hi Incoming traffic and CPU up to 100%

block ntp connections from internet, possibly you are relaying an attack

Mon Aug 17, 2015 8:55 am

We are talking about the upgrade menu. Not "Release Switch menu".

If you are on 6.30, "Bugfix" should bring Updates to 6.30.1, 6.30.2, 6.30.3
If you are on 6.31, "Bugfix" should bring Updates to 6.31.1, 6.31.2, 6.31.3

Not downgrades. Even if you download the "downgrade" and reboot, it will not be installed as it is an older version than the installed one. You will have to run /system package downgrade additionally. So in my opinion this feature is not here to switch between releases, but for changing update channel.

No, the menu provides ability to downgrade as well. So everything works as it should. If one of the "new" features in v6.31 is causing you problems, you can "downgrade" to the last stable v6.30.3

Mon Aug 17, 2015 9:12 am

Anyone with Reboots happening, please make supout.rif file after the reboot and email it to support@mikrotik.com
Thank you!

npero · Mon Aug 17, 2015 10:08 am

Send it yesterday, ticket#2015081666000091, 6.31 kill 750UP after sometime best case reset by watchdog or just hung up

, disabling wireless package also not working it is end up by kernel failure.
SXT device and BaseBox seem to me not affected, lost one Groove52 but not have access to him to see what happens.

Łukasz · Mon Aug 17, 2015 10:41 am

Send it yesterday, ticket#2015081666000091, 6.31 kill 750UP after sometime best case reset by watchdog or just hung up , disabling wireless package also not working it is end up by kernel failure.
SXT device and BaseBox seem to me not affected, lost one Groove52 but not have access to him to see what happens.

I upgrade 2 omnitiks UPA and with both is problem - lost contact with devices (only ping answer). I don't upgrade rest of omnitiks UPA.
Maybe problem is with poe-out devices?
Mikrotik, please check it.

You promised me that [Ticket#2015070966000221] will be fix in this relase - it is still not fixed.

patrick7 · Mon Aug 17, 2015 11:41 am

@normis OK, I understand. But then you should propably consider renaming this menu. Commands like update, check-for-updates for downgrades is a little bit confusing

Mon Aug 17, 2015 12:24 pm

If one of the "new" features in v6.31 is causing you problems, you can "downgrade" to the last stable v6.30.3

6.30.3? Has it been released already?

NetworkMeister · Mon Aug 17, 2015 12:42 pm

Could not add VirtualAP via winbox on my RB751G-2HnD, "L2 MTU" field is empty and grey (read only) so i get error message: "Couldn't add New Interface - l2mtu wrong (6)". It is only possible to add interface via terminal. I tried with winbox 2.2.18 and 3.0rc12.
On my second router RB2011 I can normally add VirtualAP via winbox.

I confirm this bug. Same problem on RB951G-2HnD, tried on 2 different routers. Works OK on 6.30.2, but throws "Couldn't add New Interface - l2mtu wrong (6)" on 6.31.

trottolino1970 · Mon Aug 17, 2015 12:44 pm

Could not add VirtualAP via winbox on my RB751G-2HnD, "L2 MTU" field is empty and grey (read only) so i get error message: "Couldn't add New Interface - l2mtu wrong (6)". It is only possible to add interface via terminal. I tried with winbox 2.2.18 and 3.0rc12.
On my second router RB2011 I can normally add VirtualAP via winbox.
I confirm this bug. Same problem on RB951G-2HnD, tried on 2 different routers. Works OK on 6.30.2, but throws "Couldn't add New Interface - l2mtu wrong (6)" on 6.31.

i resolve this problem with installation wireless-fp and non cm

Caci99 · Mon Aug 17, 2015 1:14 pm

please fix connection-rate problem in next bug release!

http://forum.mikrotik.com/viewtopic.php ... 12#p495212

Second this, it is not working any more. I noticed this since version 6.30 (but might have happened before).

jiaxy001 · Mon Aug 17, 2015 2:15 pm

Be very careful with this release now lost one wireless link with Groove52 not come back after update.
Also on 750UP try it to disable wireless package also lost device but they reboot after couple minutes by him self with log message kernel failure but wireless not disabled.
One 750GL every two minute in log say interface,warning ether5-slave-local excessive broadcasts/multicasts, probably a loop.

I update lot of device in last couple years this is maybe second time to lost device after update, that Groove is just simple wireless bridge but not have access in next 7 days .

My friend, I do not speak English while I'll try to make my input clear enough.

I have the same issue as you did on my RB2011UiAS-2HnD, which was working perfectly on ver. 6.29. After I upgraded it to ver. 6.31 precisely every 30 secs the log showed "ether2-master-local excessive broadcasts/multicasts, probably a loop".

I tried to reset the configuration & spent long time to sniff the imcoming packets in my LAN but found nothing.

Eventually this issue was 'resolved' by disconnecting the device (a QNAP TS-219PII NAS attached in my LAN to provide file storage). Then i connected the NAS to the empty ether5 port. It's working properly now. One day has passed since I disconnected ether2 & no more excessive broardcasts message was recorded in log.

I hope my case can helps a bit for you.

Thanks.

jiaxy001 · Mon Aug 17, 2015 2:19 pm

Be very careful with this release now lost one wireless link with Groove52 not come back after update.
Also on 750UP try it to disable wireless package also lost device but they reboot after couple minutes by him self with log message kernel failure but wireless not disabled.
One 750GL every two minute in log say interface,warning ether5-slave-local excessive broadcasts/multicasts, probably a loop.

I update lot of device in last couple years this is maybe second time to lost device after update, that Groove is just simple wireless bridge but not have access in next 7 days .

My friend, I do not speak English while I'll try to make my input clear enough.

I have the same issue as you did on my RB2011UiAS-2HnD, which was working perfectly on ver. 6.29. After I upgraded it to ver. 6.31 precisely every 30 secs the log showed "ether2-master-local excessive broadcasts/multicasts, probably a loop".

I tried to reset the configuration & spent long time to sniff the imcoming packets in my LAN but found nothing.

Eventually this issue was 'resolved' by disconnecting the device (a QNAP TS-219PII NAS attached in my LAN to provide file storage). Then i connected the NAS to the empty ether5 port. It's working properly now. One day has passed since I disconnected ether2 & no more excessive broardcasts message was recorded in log.

I hope my case can help a bit for you.

Thanks.

mistry7 · Mon Aug 17, 2015 3:50 pm

750Up Watchdog problem

#2015081766000535
I sended with autosupout!

darkprocess · Mon Aug 17, 2015 3:58 pm

Hello

I found a small bug with Winbox and this release.

I'm unable to ceeate a virtual AP using the GUI. THE L2MTU is missing and you can't put it. The field is grey.

Only solution by command line

Panoulis · Mon Aug 17, 2015 4:46 pm

hi i just updated to 6.31 from 6.30.2
and there is a bug with R11e-2HPnD cards

my setup is
Intel C2D E6400
2Gb Memory
RB14e
2x R11e-2HPnD
Intel 82571EB Gigabit Quad Copper Ethernet

with 6.30.2 i can see both R11e-2HPnD as diferent interfaces
but when i upgrade to 6.31 only on of the R11e-2HPnD card is recognized by ROS

Downgrading to 6.30.2 solves the problem and both interfaces are availiable in ROS

nescafe2002 · Mon Aug 17, 2015 5:37 pm

upgraded to 6.31 on CRS125. IP phones cannot connect to gateway after CRS125 reboot. Had to manually restart each phone. Not acceptable so reverted to 6.30.2. All fine again.

Could you try removing the UDP connections to you SIP server from the conntrack table ( IP > Firewall > Connections )?

mmabob · Mon Aug 17, 2015 10:25 pm

After updating to 6.31 on a pair of RB750 routers, I now have Issues with the ovpn connection between them.

The branch office router that connects to the main router will connect enough that I can send ping traffic over it, but as soon as I try to use the connection for useful data such as connecting to a computer on the other side of the connection, I see in the log "ovpn-out1: terminating... - TLS failed"

I don't know if it is the slightly larger amount of data being transferred that is relating to the issue or what the problem is. It worked well on version 6.27 which we were running before this.

Is anyone else seeing this?

I appreciate any of your help. I currently have a branch office without a useful connection.

I downgraded the routers back to 6.27 by copying the routeros-mipsbe-6.27.npk file to the routers and selecting System Package Downgrade. After the reboot on the older RouterOS, the OpenVPN links are now working properly again.

MAYAKNET · Mon Aug 17, 2015 10:53 pm

Пишу второй раз. После обновления до релиза 6.31, на HP-сервере, с установленным RouterOS для X86 возникла такая проблема. Вместо 32 ядер сервера - осталось только 1 ядро. И загрузка процессора показывало 100%.

pbr3000 · Tue Aug 18, 2015 5:12 am

Send it yesterday, ticket#2015081666000091, 6.31 kill 750UP after sometime best case reset by watchdog or just hung up , disabling wireless package also not working it is end up by kernel failure.
SXT device and BaseBox seem to me not affected, lost one Groove52 but not have access to him to see what happens.
I upgrade 2 omnitiks UPA and with both is problem - lost contact with devices (only ping answer). I don't upgrade rest of omnitiks UPA.
Maybe problem is with poe-out devices?
Mikrotik, please check it.

You promised me that [Ticket#2015070966000221] will be fix in this relase - it is still not fixed.

We had the same issue on some RB450g. Just pings and no ssh, http etc. Reverted to 6.30.2.
PS: v6.31 worked fine for some time, like 12 hours or so. I don't sent the supout file because there is no access to do it.

pbr3000 · Tue Aug 18, 2015 5:17 am

If one of the "new" features in v6.31 is causing you problems, you can "downgrade" to the last stable v6.30.3
6.30.3? Has it been released already?

I guess no. It was just an example. Normis, can you confirm this?

infused · Tue Aug 18, 2015 6:09 am

Virtual Support... oh hell yeah.

makstex · Tue Aug 18, 2015 9:02 am

RB750:
Switch DES-1005P on any port mikrotik.
Auto Negatiation is ON.
Link at intervals of 2 seconds up...down...up....down...
Downgrade to 6.30.2 solved a problem,

Reborns · Tue Aug 18, 2015 9:38 am

RB750:
Switch DES-1005P on any port mikrotik.
Auto Negatiation is ON.
Link at intervals of 2 seconds up...down...up....down...
Downgrade to 6.30.2 solved a problem,

Yes i can confirm that ..

phuang3 · Tue Aug 18, 2015 10:01 am

Problem with 6.31 on NetMetal 5. After reboot, I can see wireless-6.31-misbe.npk was still there and Netbox showed 6.30! Downgrade to 6.30.2 successfully, but the Netbox is showing 6.31... I got confused!

After upgrade to 6.31

After downgrade to 6.30.2. Still showing 6.31

kostas22 · Tue Aug 18, 2015 10:22 am

After updating to 6.31 on a pair of RB750 routers, I now have Issues with the ovpn connection between them.

The branch office router that connects to the main router will connect enough that I can send ping traffic over it, but as soon as I try to use the connection for useful data such as connecting to a computer on the other side of the connection, I see in the log "ovpn-out1: terminating... - TLS failed"

I don't know if it is the slightly larger amount of data being transferred that is relating to the issue or what the problem is. It worked well on version 6.27 which we were running before this.

Is anyone else seeing this?

I appreciate any of your help. I currently have a branch office without a useful connection.

I also have issues when updating RB750 routers from 6.23 to 6.31. After the updating I don't have internet, if I downgrade back to 6.23 is fine.

demonster · Tue Aug 18, 2015 11:01 am

I have issue with switch port flapping with 6.31 on "model: 750", after downgrade to 6.30.2 all is ok

I have same problem on two RB750. After downgrade to 6.30.2 port flapping reveals on week or more - reboot needed.

WirelessRudy · Tue Aug 18, 2015 11:16 am

Finally it's great to know that the CHR is now available...

CCR= "Cloud Core Router", what stands "CHR" for? Cloud Host router? What new device would this be anyway?

Tue Aug 18, 2015 11:17 am

CHR is a virtual machine image, it is not a device.

infused · Tue Aug 18, 2015 11:26 am

So out of interest, is virtual routers going to ever come back to Mikrotik in the form of multi core routers? I totally dig vmware images, but that'd be nice too.

Tue Aug 18, 2015 11:31 am

Multicore and 64Bit is already supported by CHR image. CHR is not only for VMware, it is for any virtual system

infused · Tue Aug 18, 2015 11:46 am

I mean inside of the Mikrotik OS, to spawn up virtual routers within RouterOS, like metarouter, which works on CCR etc. We have a couple of Fortinets to do this where we need it.

Cha0s · Tue Aug 18, 2015 11:48 am

I mean inside of the Mikrotik OS, to spawn up virtual routers within RouterOS, like metarouter, which works on CCR etc. We have a couple of Fortinets to do this where we need it.

CCRs do not support Metarouter.

Tue Aug 18, 2015 11:52 am

That is a different topic, CHR works ourside the RouterBOARD, either in your PC or in the cloud or somewhere.
RouterOS built in virtualization also will be re-made, but nothing to announce yet.

angboontiong · Tue Aug 18, 2015 12:12 pm

Normunds,

Can you explain more detail on this 3:
*) fixed ethernet stopping on RB NetMetal / SXTG-5HPacD 10Mbit and 100Mbit links
-can you share more detail about this as i believe there is some case study.

*) wireless - implemented l2mtu update if wireless-cm2 is enabled
*) wireless - improved WMM-PowerSave support in wireless-cm2 package
-What's the different fo the cm2 and fp package? which is more faster on the packet pass-through?
-It's the cm2 compatible with fp package or they only can talk with the same package running?

peter37577 · Tue Aug 18, 2015 12:21 pm

working fine on several devices (CCR, CRS, Groove, 951G, 751G, hap-lite)

one common issue
/system shutdown via Winbox and/or console, ssh, telnet is not working anymore
The device does make a reboot instead.
This issue is found on the platforms mips-be and smips (tilera not tested but likely also affected).
Edit Issue is also present on ppc platform (RB1200 and RB1100AH tested).
The issue is gone after downgrading to 6.30.2. Edit end

I hope this is a bug. Or is shutdown not supported anymore?

The same observed on 751, 951 and 2011 with mipsbe.

angboontiong · Tue Aug 18, 2015 12:48 pm

working fine on several devices (CCR, CRS, Groove, 951G, 751G, hap-lite)

one common issue
/system shutdown via Winbox and/or console, ssh, telnet is not working anymore
The device does make a reboot instead.
This issue is found on the platforms mips-be and smips (tilera not tested but likely also affected).
Edit Issue is also present on ppc platform (RB1200 and RB1100AH tested).
The issue is gone after downgrading to 6.30.2. Edit end

I hope this is a bug. Or is shutdown not supported anymore?
The same observed on 751, 951 and 2011 with mipsbe.

Shutdown is not the safe option, it should be removed for security purposes.

Chupaka · Tue Aug 18, 2015 12:54 pm

Вместо 32 ядер сервера - осталось только 1 ядро.

I can confirm: 6.31on x86 detects only 1 CPU core:

[admin@TestPlace] > sys hardware pr
  multi-cpu: yes
[admin@TestPlace] > sys resource pr
                cpu-count: 1

if you check 'Allow x86-64' in WinBox under System -> Resources -> Hardware, then all CPUs are back. what's Terminal command for 'Allow x86-64'? Can't find anywhere...

WirelessRudy · Tue Aug 18, 2015 1:13 pm

750Up Watchdog problem

#2015081766000535
I sended with autosupout!

I can confirm. Yesterday updated one line (route) of 6 nodes towards AP fm 6.27 to 6.31. It includes 2 rb750UP's. One already rebooted 4 times for watchdog. (No ip set for watchdog.) I'll revert back to 6.30.2 because that seems to work on at least 8 other 750UP's out in the field..
It took some 12 hours for the first reboot to occur.

The other one has not been rebooted but than again, watchdog timer is also not set..... It runs now for almost 17 hours.

Tue Aug 18, 2015 1:40 pm

We have found out that there are two things which causes reboot on router with RouterOS 6.31. Routers with PoE-OUT are rebooting themselves and also routers which has MetaROUTER.

Is there any other router rebooting itself with 6.31 which does not have PoE-OUT or MetaROUTER on it?

infused · Tue Aug 18, 2015 1:46 pm

I mean inside of the Mikrotik OS, to spawn up virtual routers within RouterOS, like metarouter, which works on CCR etc. We have a couple of Fortinets to do this where we need it.
CCRs do not support Metarouter.

That's my point.

ufm · Tue Aug 18, 2015 1:53 pm

That is a different topic, CHR works ourside the RouterBOARD, either in your PC or in the cloud or somewhere.
RouterOS built in virtualization also will be re-made, but nothing to announce yet.

As I understand - CHR without limitations and 64 version for bare metal x86 - not planned?

Tue Aug 18, 2015 2:02 pm

As I understand - CHR without limitations

planned, but as a subscription service

64 version for bare metal x86 - not planned?

unclear, but most likely CHR will get priority with all 64Bit and multicore support for now

Chupaka · Tue Aug 18, 2015 2:27 pm

64 version for bare metal x86 - not planned?
unclear, but most likely CHR will get priority with all 64Bit and multicore support for now

so, what is 'Allow x86-64' tick in x86 RouterOS?

Tue Aug 18, 2015 2:54 pm

Chupaka: Bug, on X86 regular this should not be shown. This is only for CHR image

rajo · Tue Aug 18, 2015 3:09 pm

We have found out that there are two things which causes reboot on router with RouterOS 6.31. Routers with PoE-OUT are rebooting themselves and also routers which has MetaROUTER.

Is there any other router rebooting itself with 6.31 which does not have PoE-OUT or MetaROUTER on it?

I don't have either and I get reboots on one of my RB450G routers, if I have 6.31 loaded.

marosi · Tue Aug 18, 2015 3:18 pm

Found something strange. On wireless links with NV2 and a secret. If You change something on the link or reboot the device then the link goes down every 3-5 seconds with "key exchange timeout" in the log.

When i type in the NV2 secret by hand on the client of the wireless link and klick "apply", the link comes back and stay stable. So do not reboot or change anything or you have to type the NV2 secret and klick apply. Very strange.

I tested this about 10 times on different SXTs (2.4 and 5Ghz) to confirm this behavior.

WirelessRudy · Tue Aug 18, 2015 3:29 pm

Found something strange. On wireless links with NV2 and a secret. If You change something on the link or reboot the device then the link goes down every 3-5 seconds with "key exchange timeout" in the log.

When i type in the NV2 secret by hand on the client of the wireless link and klick "apply", the link comes back and stay stable. So do not reboot or change anything or you have to type the NV2 secret and klick apply. Very strange.

I tested this about 10 times on different SXTs (2.4 and 5Ghz) to confirm this behavior.

I can't repeat this issue.
After I upgrade several SXT's and they come back up I also upgrade the firmware, which needs another reboot. After that units come back up and stay up. (NV2 + secret enable CPE/station).

To test again I logged into an already upgraded SXT and changed something in the config (just a scan setting in wireless) but the unit just kept on working fine.
Than I made a reboot of the SXT and after the disconnection for a while it came backup and stayed connected.

So the prescribed issue of the "key exchange timeout" is not happening on my units.

And how can you 'type' in the secret again? Has it been erased? Do you have access from the AP to the cpe? Or do you have to go to the client to login and correct?

WirelessRudy · Tue Aug 18, 2015 3:33 pm

We have found out that there are two things which causes reboot on router with RouterOS 6.31. Routers with PoE-OUT are rebooting themselves and also routers which has MetaROUTER.

Is there any other router rebooting itself with 6.31 which does not have PoE-OUT or MetaROUTER on it?

I have two rb750UP's updated to 6.31.
One with watchdog function enabled (no address) did reboot several times due the watchdog function according the log
The other one, that does NOT have the watchdog function enable is just up since the upgrade. That's 20 hours ago now...

Maybe this is a little more help in troubleshooting the issue?

WirelessRudy · Tue Aug 18, 2015 3:35 pm

CHR is a virtual machine image, it is not a device.

So its actually a "CGR", Cloud Ghost Router!

ufm · Tue Aug 18, 2015 3:44 pm

Chupaka: Bug, on X86 regular this should not be shown. This is only for CHR image

Wait-wait. You want to say - in 6.31 broken support multicore cpu on x86 and this is not a bug, it's a feature?????

Tue Aug 18, 2015 3:47 pm

The specific checkbox/setting is not supposed to be shown in PC, only in CHR image.

ufm · Tue Aug 18, 2015 3:58 pm

The specific checkbox/setting is not supposed to be shown in PC, only in CHR image.

But this checkbox enable multicore support for x86 on 6.31
Disabled multicore support - it's a bug?

Tue Aug 18, 2015 4:03 pm

The multicore support is enabled automatically. You can disable it. But in this case, there is a bug, like I said.

kotsius · Tue Aug 18, 2015 4:08 pm

i try upgrade my RB493aH .. run 5 min and blocked, cant acess via mac or ip..
power off, power on.. again 3min, again 20 min.. WTF with 6.31.. with problems i return to 6.29.1, i have lucky that these mtik was on my house, so i can unplug power, but we have some network in the city, not good option.

Problem confirmed. I upgraded 28 routers (incl. two x86 machines) from v6.29.1 to v6.31 and, on 2 of those, I completely lose Winbox access via IP or MAC, as well as SSH, a short while after each and every reboot. Otherwise, they seem to function fine.

In our case, what sets these two 951G-2HnD routers apart from the others, is that 4 of their ports are on a bridge and ether5 is used to connect to another router – an identical 951G-2HnD which serves as the former's backup, and which can be accessed without problems. When ether5 is disconnected, the problem seems to go away. Do I need to contact support?

105547111 · Tue Aug 18, 2015 4:22 pm

We have found out that there are two things which causes reboot on router with RouterOS 6.31. Routers with PoE-OUT are rebooting themselves and also routers which has MetaROUTER.

Is there any other router rebooting itself with 6.31 which does not have PoE-OUT or MetaROUTER on it?

I've got no POE or MetaROUTER and ALL the following devices uptime 3d 20:50 and floorless.@ 6.31

CCR1016-12G
CRS125-24G-RM
RB951G
RB751G
RB751U
RB9512N
RB532A

gwartass · Tue Aug 18, 2015 5:29 pm

We have found out that there are two things which causes reboot on router with RouterOS 6.31. Routers with PoE-OUT are rebooting themselves and also routers which has MetaROUTER.

Is there any other router rebooting itself with 6.31 which does not have PoE-OUT or MetaROUTER on it?

Yes, RB433UAH [Ticket#2015081766000357]
Downgrade to 6.30.2 helps.

mmabob · Tue Aug 18, 2015 7:25 pm

After updating to 6.31 on a pair of RB750 routers, I now have Issues with the ovpn connection between them.

The branch office router that connects to the main router will connect enough that I can send ping traffic over it, but as soon as I try to use the connection for useful data such as connecting to a computer on the other side of the connection, I see in the log "ovpn-out1: terminating... - TLS failed"

I don't know if it is the slightly larger amount of data being transferred that is relating to the issue or what the problem is. It worked well on version 6.27 which we were running before this.

Is anyone else seeing this?

I appreciate any of your help. I currently have a branch office without a useful connection.
I downgraded the routers back to 6.27 by copying the routeros-mipsbe-6.27.npk file to the routers and selecting System Package Downgrade. After the reboot on the older RouterOS, the OpenVPN links are now working properly again.

I may have spoke too soon. Even though the routers are now able to connect using 6.27, there still seems to be a remaining problem from the update. Now running 6.27, the ovpn connection repeatedly goes down throughout the day with "Terminating... - Peer Disconnected" in the logs. These routers had been rock solid, but now they are continuing to cause issues, and I can't say that anything else has changed other than they ran 6.31 for a day or so.

I'm not sure what is going on here, but I hope we can find some answers soon. MikroTik has been my "go to router" for any install because of it's rock solid stability. I understand mistakes can happen, but it's kind of frustrating to get burned by a firmware update.

Thanks again for your help,
Bob

Tue Aug 18, 2015 8:15 pm

Not working on my 450g, it restarts every few minutes.
On second router 2011 works fine.

Have you update the bios to?

travisms · Tue Aug 18, 2015 8:20 pm

I might have a possbile bug... I know that there was a bug that has been addressed with the sierra MC8705 lte card where the interface would rename from lte1 to lte2 upon update or backup restore.

I just experienced something similar and logged ticket #2015081866000641. I reset our hAP lite to defaults, upgraded it from 6.25 to 6.31 and then restored a backup that I made from another hAP lite. What I noticed is that the wlan1 interface renamed to wlan2 and was marked as disabled and none of the settings from the backup took effect on the wlan section.

Is this something that I am doing wrong possibly ?

HWTest · Tue Aug 18, 2015 8:37 pm

The "CCP lost compression got out of sync" L2TP/IPSec VPN client bug is still present in this version, even when compression is disabled in the ppp profile.
The "dial on demand" option in SSTP VPN client still acts strange - connects and disconnects in cca 1 second pace.
Ticket #2015081866000551

zervan · Tue Aug 18, 2015 10:40 pm

I have issue with switch port flapping with 6.31 on "model: 750", after downgrade to 6.30.2 all is ok
I have same problem on two RB750. After downgrade to 6.30.2 port flapping reveals on week or more - reboot needed.

Guys, sorry I didn't believe you - because RB751U and RB751G (and many other models) were working fine for me. I also have one old RB750, I tried to upgrade to 6.31 and it is horrible, see the log:

rb750-ros-6.31-port-flapping.png

Downgrade to 6.30.2 and everything is OK.

Mikrotik staff - you should really check this.

marosi · Wed Aug 19, 2015 2:11 am

Found something strange. On wireless links with NV2 and a secret. If You change something on the link or reboot the device then the link goes down every 3-5 seconds with "key exchange timeout" in the log.

When i type in the NV2 secret by hand on the client of the wireless link and klick "apply", the link comes back and stay stable. So do not reboot or change anything or you have to type the NV2 secret and klick apply. Very strange.

I tested this about 10 times on different SXTs (2.4 and 5Ghz) to confirm this behavior.
I can't repeat this issue.
After I upgrade several SXT's and they come back up I also upgrade the firmware, which needs another reboot. After that units come back up and stay up. (NV2 + secret enable CPE/station).

To test again I logged into an already upgraded SXT and changed something in the config (just a scan setting in wireless) but the unit just kept on working fine.
Than I made a reboot of the SXT and after the disconnection for a while it came backup and stayed connected.

So the prescribed issue of the "key exchange timeout" is not happening on my units.

And how can you 'type' in the secret again? Has it been erased? Do you have access from the AP to the cpe? Or do you have to go to the client to login and correct?

I must log in via ethernet, open the config window and nv2 tab and type the nv2 secret. It was empty.
The export of the interface settings gave no nv2 secret.

docmarius · Wed Aug 19, 2015 8:02 am

We have found out that there are two things which causes reboot on router with RouterOS 6.31. Routers with PoE-OUT are rebooting themselves and also routers which has MetaROUTER.

Is there any other router rebooting itself with 6.31 which does not have PoE-OUT or MetaROUTER on it?

OmniTIK UPA-5HnD - PoE out is disabled, but reboot occurs at some 5 min. Difference is that RB is accessible after reboot (back to 6.30.2 fixes the issue).

Wed Aug 19, 2015 9:43 am

We have found out that there are two things which causes reboot on router with RouterOS 6.31. Routers with PoE-OUT are rebooting themselves and also routers which has MetaROUTER.

Is there any other router rebooting itself with 6.31 which does not have PoE-OUT or MetaROUTER on it?
OmniTIK UPA-5HnD - PoE out is disabled, but reboot occurs at some 5 min. Difference is that RB is accessible after reboot (back to 6.30.2 fixes the issue).

Have you update the bios to?

CyberTod · Wed Aug 19, 2015 9:53 am

Problem confirmed. I upgraded 28 routers (incl. two x86 machines) from v6.29.1 to v6.31 and, on 2 of those, I completely lose Winbox access via IP or MAC, as well as SSH, a short while after each and every reboot. Otherwise, they seem to function fine.

Same problem. So far upgraded only 6 devices. But one of them has this problem. Actually as it was a new device, just bought, upgraded and installed I spent hours trying to log in and wandering what I did in firewall to cut myself off. A port scan shows that the ports are open - 22,23,80,8291. A tcpdump shows that I get responses to my requests but telnet and ssh sessions timeout on login. A winbox session stays on 'logging in' indefinately. Webfig opens the login page after 1-2min and after trying to login and waiting for 1-2min more says 'internal server error'. Mac-telnet fails also. I haven't configured all so Romon is not enabled and cannot say if it will work. The only management tool that work is snmp, but I don't have the practice to enable write access so I cannot reboot from there.

docmarius · Wed Aug 19, 2015 10:44 am

Have you update the bios to?

I have a RB450G on which I updated the bootloader to 3.24 with 6.30.2. After downgrade to 6.29.1, I had to activate the backup bootloader to be able to start it and restore the older 3.22 bootloader. So that was a lesson learned.

Now, would you risk upgrading the bootloader on a device that randomly resets and brick the device on upgrade or downgrade, with that device mounted on a hardly accessible rooftop, or would you choose to stay with the previous stable version of ROS

?

abis · Wed Aug 19, 2015 10:59 am

951Ui-2HnD

system reboot if o chose Shutdown, from WinBox and from command line. I'll update the post for other models...

Wed Aug 19, 2015 11:52 am

951Ui-2HnD

system reboot if o chose Shutdown, from WinBox and from command line. I'll update the post for other models...

Have you update the bios to?
See on my website, in Dutch but folow the printscreens a made
http://www.wirelessinfo.be/index.php/mi ... /rbupgrade

abis · Wed Aug 19, 2015 12:03 pm

951Ui-2HnD

system reboot if o chose Shutdown, from WinBox and from command line. I'll update the post for other models...
Have you update the bios to?
See on my website, in Dutch but folow the printscreens a made
http://www.wirelessinfo.be/index.php/mi ... /rbupgrade

yep, is to 3.24

Awax · Wed Aug 19, 2015 12:48 pm

I have issue with switch port flapping with 6.31 on "model: 750", after downgrade to 6.30.2 all is ok
I have same problem on two RB750. After downgrade to 6.30.2 port flapping reveals on week or more - reboot needed.
Guys, sorry I didn't believe you - because RB751U and RB751G (and many other models) were working fine for me. I also have one old RB750, I tried to upgrade to 6.31 and it is horrible, see the log:
rb750-ros-6.31-port-flapping.png
Downgrade to 6.30.2 and everything is OK.

Mikrotik staff - you should really check this.

I confirm, I have the same problem with RB750

Wed Aug 19, 2015 1:09 pm

Anyone with reboot issues, please try v6.32rc1 that we just released. Let us know if this helps.

neophyte · Wed Aug 19, 2015 1:26 pm

After upgrade to 6.31 and 6.32.rc1 /interface wireless print output is empty,wifi dont work.
X86 with dlink DWA-520
Wireless (Atheros AR5213)
/system resource pci print
Atheros AR5001X+ Wireless Network Adapter (rev: 1)

WirelessRudy · Wed Aug 19, 2015 1:35 pm

Anyone with reboot issues, please try v6.32rc1 that we just released. Let us know if this helps.

And if it does would this make it a bug fix to 6.31? That would make it a v6.31.1 isn't it? Why is it a 6.32rc1 now? 6.32 is not out yet......
The new numbering works but this is confusing again.....

Wed Aug 19, 2015 2:00 pm

Anyone with reboot issues, please try v6.32rc1 that we just released. Let us know if this helps.
And if it does would this make it a bug fix to 6.31? That would make it a v6.31.1 isn't it? Why is it a 6.32rc1 now? 6.32 is not out yet......
The new numbering works but this is confusing again.....

Are you expecting v6.31.1rc1??

I think they are using 6.32rc versions to confirm fixes and later on those will be backported to 6.31.x

Wed Aug 19, 2015 2:03 pm

WIRELESSRUDY http://forum.mikrotik.com/viewtopic.php ... .1#p490862

Ansy · Wed Aug 19, 2015 2:58 pm

Nightly I've updated my RB750UP (yes, 3xPoE out) to ROS 6.32/fw3.24, I was afraid but I really need in Use IP Firewall on Bridge + Simple Queues working there.

Well done MT! Finally you (I hope, crossing my fingers... ) fixed it!

[Ticket#2015080366000534] http://forum.mikrotik.com/viewtopic.php ... 95#p493595

16+ hours uptime, night workload & day workload, up to [70% CPU, 25Mbit/s total, 7k+ firewall connections]... and IT WORKS AND WORKS FOR NOW!

Process management doesn't eat CPU, no weird log messages, port forwarding OK, Simple Queues shape traffic, even Graphing is switched on and collects info.

I'm very very glad. Will monitor and tell you later. Thanks!

abis · Wed Aug 19, 2015 4:36 pm

up and running, about four days, after upgrade to 6.31 on that models:

rajo · Wed Aug 19, 2015 5:40 pm

Anyone with reboot issues, please try v6.32rc1 that we just released. Let us know if this helps.

I obviously replied too soon. Got a lockup reboot on an RB450G with v6.32rc1. This time it took about 50 minutes for it to happen. Going back to 6.30.2

zm0m · Wed Aug 19, 2015 6:23 pm

RB951G-2HnD
UPD: RB1100AHx2

In WinBox in Status tab of OVPN client(or server) interface are not displayed IP-values of Local Address and Remote Address. The connection is established and is operating normally.

myke1124 · Wed Aug 19, 2015 7:29 pm

RB411AH ROS 6.31 firmware 3.24

/export

is broken. Terminal appears to be locked up. [CTRL]+C does not stop the process. Opening a new terminal window is also locked up. Unable to reboot device via winbox.
RB Metal 5SHPN with same firmware works.
hAP lite works.

Sent supout.rif

alexspils · Wed Aug 19, 2015 7:46 pm

Hello!
I have multicore x86 router.
With 6.30.2 it shows 4 cores and all ok
When i installed 6.31 or 6.32rc1 it shows only 1 core.
Soluton to resolve:enable checkbox Allow x86-64 and reboot router.
After that routers uses all 4 cores.And yes, with enabled checkbox Allow x86-64 system resource print shows 4gb ram
uptime: 11m34s
version: 6.31
build-time: Aug/14/2015 15:42:51
free-memory: 3670.9MiB
total-memory: 3945.6MiB
cpu: Intel(R)
cpu-count: 4
cpu-frequency: 3093MHz
cpu-load: 38%
free-hdd-space: 818.6MiB
total-hdd-space: 949.6MiB
write-sect-since-reboot: 42624
write-sect-total: 42624
architecture-name: x86
board-name: x86
platform: MikroTik

Mikrotik, may be leave this option in x86 version?

JB172 · Wed Aug 19, 2015 7:49 pm

RB411AH ROS 6.31 firmware 3.24
Code: Select all
/export
is broken. Terminal appears to be locked up. [CTRL]+C does not stop the process. Opening a new terminal window is also locked up. Unable to reboot device via winbox.
RB Metal 5SHPN with same firmware works.
hAP lite works.

Sent supout.rif

I install ROS 6.31 firmware 3.24 on RB411AH and all working fine. Export is working also.

ericdupre · Wed Aug 19, 2015 9:25 pm

I got one error from my CCR1009-8G-1S-1S+PC. When I check for new update, I saw the last update version but I can't retrieve the change log and I can't follow the upgrade.

getting change log / ERROR: connection time out.

So, I need to do the manual upgrade.

Any one's having this issue?

I saw other posted where the issue was fix with some configuration change like DNS and FIREWALL. I tested that on my CCR1009-8G-1S-1S+PC whit no go.

Update Error.PNG

saaremaa · Wed Aug 19, 2015 10:23 pm

Similar problem with multicore x86 platform.
Tonight we have updated to version 6.31 x86 server (8 core) with RouterOS and have stopped working 7 cores. Only 1 core was used with utilization 80-95%.
Big thanks alexspils for your solution. After that routers uses all 8 cores and cpu-load between 4% and 7 %

uptime: 8m18s
version: 6.31
build-time: Aug/14/2015 15:42:51
free-memory: 3543.8MiB
total-memory: 3674.9MiB
cpu: Intel(R)
cpu-count: 8
cpu-frequency: 3392MHz
cpu-load: 4%
free-hdd-space: 54.8GiB
total-hdd-space: 55.0GiB
write-sect-since-reboot: 1434
write-sect-total: 1434
architecture-name: x86
board-name: x86
platform: MikroTik

plankanater · Thu Aug 20, 2015 4:15 am

Seems to be an issue with 750up. I have upgraded two of them to 6.31 they take the upgrade but as soon as i log into them they lock up and go offline. I power cycle them and they pass traffic, then as soon as I log into them it locks up again. By lock up I mean no ping, no mac ping, no passing traffic.

Thanks,

Thu Aug 20, 2015 8:32 am

rajo - Are you using MetaROUTER?

For anyone who is still experiencing reboots with 6.32rc1 - please send supout files to support@mikrotik.com

Remember that file must be generated after device has rebooted itself but before you reboot it manually or do a power cycle on it.

AdminioBanderas · Thu Aug 20, 2015 9:33 am

Anyone with Reboots happening, please make supout.rif file after the reboot and email it to support@mikrotik.com
Thank you!

Just send supout.rif.
Please do not make experiments on live people.

RB751G-2HnD v6.31 / 3.24 (updated to 6.32rc1 yesterday) reboots problem by watchdog.

Hi guys!
6.32rc1 didnt solve the problem. Very sad (( And very bad. Cant work normally for a 4 days!
Please help. Is there archive of old ROS? Need 6.29.

If turn watchdog off - router just hung up and must switch it off/on manually.
Pay attention, before watchdog problem there are strange error "excessive broadcasts/multicast, probably a loop" in log Sometimes on ether5, sometimes on ether1. But now connected only ISP to port 1 and my notebook to port 5. There is cannot be any loops.

Thanx.

(Only one of two rb751g has watchdog problem in v6.31. But have up to 10 mikrotiks. Some 751 and 951. And after v.6.31 often cannot login to diffirent routers by winbox until reboot router manually. Just see "Logging in..." forever)

michaelcarey · Thu Aug 20, 2015 9:43 am

I confirm, I have the same problem with RB750

And me too... my RB750 had interfaces going up and down after upgrading to v6.31.

Everything fine after going back to v6.30.2

Thu Aug 20, 2015 11:15 am

This looks to be quite dangerous version. It killed one of my testing devices. I still wonder why all the people massively install every new version without any testing in the laboratory before.

Thu Aug 20, 2015 11:30 am

This looks to be quite dangerous version. It killed one of my testing devices. I still wonder why all the people massively install every new version without any testing in the laboratory before.

agreed, especially when you still have nice bugfix version to keep on those production routers.

on topic: v6.32rc1 solved my watchdog reboot problems on Omnitik UPA.

Thu Aug 20, 2015 1:27 pm

Issue where only one CPU core was used on x86 devices starting from 6.31 should be fixed in 6.32rc2. For those of you who had this problem, please test this fix, if it is possible to use rc version in your environment.

JB172 · Thu Aug 20, 2015 1:40 pm

Issue where only one CPU core was used on x86 devices starting from 6.31 should be fixed in 6.32rc2. For those of you who had this problem, please test this fix, if it is possible to use rc version in your environment.

Please change in download page in "Please select version:" Release Candidate (6.32rc1) to (6.32rc2)
The links for downloads shows right for 6.32rc2

Thu Aug 20, 2015 1:47 pm

It has not changed yet. You have to wait for a while (cached entry).

plankanater · Thu Aug 20, 2015 2:28 pm

Seems to be an issue with 750up. I have upgraded two of them to 6.31 they take the upgrade but as soon as i log into them they lock up and go offline. I power cycle them and they pass traffic, then as soon as I log into them it locks up again. By lock up I mean no ping, no mac ping, no passing traffic.

Thanks,

Both devices I had lock up on me were getting POE in port one and putting POE out on other ports.

rajo · Thu Aug 20, 2015 2:58 pm

rajo - Are you using MetaROUTER?

No. I am not using MetaROUTER.

I did create a supout; however, I am very reluctant to send it, because (if a third party got hold of it), it could allow access into our internal network, since the supout generation process does not randomize password and encryption secrets/keys.

If I can find time between now and sometime next week to take the network offline, randomize the sensitive info, and reproduce the lockup/reboot, I will generate a supout and send it then.

alexspils · Thu Aug 20, 2015 4:23 pm

yes, 6.32rc2 fixes multicore support, but without checkbox allow x86-64 router uses only 2gb of ram...Normis, may be You can leave this option for x86 systems with more than 2 gb of ram?

schmitt330 · Thu Aug 20, 2015 4:46 pm

Could not add VirtualAP via winbox on my RB751G-2HnD, "L2 MTU" field is empty and grey (read only) so i get error message: "Couldn't add New Interface - l2mtu wrong (6)". It is only possible to add interface via terminal. I tried with winbox 2.2.18 and 3.0rc12.
On my second router RB2011 I can normally add VirtualAP via winbox.

I am having the same problem. Wasn't sure if it was something that maybe we weren't supposed to do from the beginning and they fixed it so we can't by mistake, lol.

rextended · Thu Aug 20, 2015 7:14 pm

WARNING: BUG: Password fields are replaced with real ****** when on winbox are clicked OK or apply!!!

Temporary solution: work with hide password unchecked....

The problem, probably is related to:
*) winbox - also hide passphrase in CAPsMAN with "Hide Password"

Geeek · Thu Aug 20, 2015 8:46 pm

We did update for 6:31 and several bondings stopped working .

pbr3000 · Thu Aug 20, 2015 9:35 pm

This looks to be quite dangerous version. It killed one of my testing devices. I still wonder why all the people massively install every new version without any testing in the laboratory before.
agreed, especially when you still have nice bugfix version to keep on those production routers.

on topic: v6.32rc1 solved my watchdog reboot problems on Omnitik UPA.

+1

It killed two of my testing devices (RB450g) and I could not to generate the supout file. CCR (1016) is working fine until now, but I am not using v6.31 in any production device yet. 6.30.2 is very stable on my network scenario.

JonasJ · Thu Aug 20, 2015 9:37 pm

i try upgrade my RB493aH .. run 5 min and blocked, cant acess via mac or ip..
power off, power on.. again 3min, again 20 min.. WTF with 6.31.. with problems i return to 6.29.1, i have lucky that these mtik was on my house, so i can unplug power, but we have some network in the city, not good option.

Upgraded two RB493G, one freezes after 40 - 60 minutes of use, no IP access, no MAC access, only Power off power on to restore...
Another - still running OK for 1 day 6 hours.

Difference in between these two routers was presence of Openflow package on the failing router. Disabled it and more than one hour still running.

Thu Aug 20, 2015 11:11 pm

Ok, I owe the description what happend when I updated RB532A+564 to 6.31.

Used automatic download and update. Everything looked normally. After the reboot the device lost its config, so I had to connect to it by mac address and with default admin/blank password. Checked the packages and realised that only system was installed. So I pushed the relevant package files (advanced, ppp, dhcp, security, wirelesscm2 and system) again into the flash. After the transfer was finished, I let the router to reboot. After that, the router went into boot loop.

Hung a serial cable to it.

When I entered the booting menu to check what was set, I have found that the option: "Boot from compact flash only" was selected.
And I would like to know: Why??

I said, fine, maybe just some strange thing, and set the nand booting hoping that eventually the system is fine.

Loading kernel from nand... ok.
setting up elf image... ok.
junping to kernel code.
Error: no system package found!
Kernel panic - not syncing: Attempted to kill init!

and reboot, reboot, reboot...

So, it was clear that it is not fine at all. The time for netinstall. Have done the memory test just to be sure nothing was roasted before the netinstall. No errors.

I have downloaded the netinstall from mipsle/6.30.2. Did everything correctly (firewall, antivirus, admin, fixed ip....), set the booting from ethernet, the device appeared in the netinstall. Selected the wanted packages and pressed "Install". Some "preparation" or what (do not remember exactly) was displayed, but the bar did not started and nothing was uploaded. Tried several times, no luck.

So I digged some old netinstall (6.24) from my archive and used it to upload 6.30.2 files. What a miracle! It worked unlike the actually downloaded netinstall.

Now, the device is back again and prepaired to serve further.

Mikrotik, please, as it was said here hundreds times: Do not push the untested beta versions as "stable" to public. Few previous versions was quite fine for almost all people, but after some time it looks again that 6.31 is really something that you were not tested pretty much.

I wish Mikrotik could publish their testing manual / prescription and let the public to suggest its extensions / additions. Because it looks to be obvious that the inhouse testing is not made as it should be before release.

Also, please, do not fool us that 6.32rcXX is something that would be converted in 6.30.X. Please, keep the versioning, so, if you maintain 6.30 for bugfixes, make the 6.30rcXX and then the correct 6.30.X but keep it distinct from 6.31rcXX and from 6.32rcXX. Do not mix the bugfixing with new features release, so we will never saw sentence like "This will not be corrected in 6.30.X but in 6.31..." Everything should be corrected in the version actually maintained for bugfixes and not delayed to the next version. The subsequent version should always adopt the bugfixes from previous version and not to bring the old bugs again. Otherwise your new huge versioning system is useless.

Hope I wrote it in the uderstandable way. I am really tired now...

pbr3000 · Fri Aug 21, 2015 12:28 am

i try upgrade my RB493aH .. run 5 min and blocked, cant acess via mac or ip..
power off, power on.. again 3min, again 20 min.. WTF with 6.31.. with problems i return to 6.29.1, i have lucky that these mtik was on my house, so i can unplug power, but we have some network in the city, not good option.
Upgraded two RB493G, one freezes after 40 - 60 minutes of use, no IP access, no MAC access, only Power off power on to restore...
Another - still running OK for 1 day 6 hours.

Difference in between these two routers was presence of Openflow package on the failing router. Disabled it and more than one hour still running.

We had this problem (or symptoms) without the openflow package installed. Tested on 6.31rc16 and 6.31. The running services are dhcp, pppoe, routing and mpls. No hotspot or special configuration. But like I said before, I can't report because there are no access to generate the supout file.

buckthis · Fri Aug 21, 2015 1:11 am

on a 2011UiAS-2HnD.

Is this a bug, configuration issue or just noise?
Not seen on previous versions.

Logs showing "tx loop on interface" ...

Time: Aug/20/2015 10:03:19
Buffer: memory
Topics: interface warning
Message: (unknown): tx loop on interface (1)

Fri Aug 21, 2015 7:48 am

yes, 6.32rc2 fixes multicore support, but without checkbox allow x86-64 router uses only 2gb of ram...Normis, may be You can leave this option for x86 systems with more than 2 gb of ram?

32-bit Kernel always was limited to 2GB of RAM, i think they eventually will add it, but as we can see from CHR there are still many things and features not working with that checkbox, so give them time to fix it.

Fri Aug 21, 2015 9:04 am

Everyone who is experiencing kernel failure with MetaROUTER installed on router - this issue will be fixed in 6.32rc3.

buckthis - It is strange that log shows problem on (unknown) interface. Is there a bridge port/s on your router which have/has been deleted? If there is, then under bridge ports menu there should be entry with interface name (unknown).

rextended - Can you please name where are you setting this password? I have tested with PPP/Secrets and User/Password. Have not managed to reproduce.

rextended · Fri Aug 21, 2015 9:14 am

rextended - Can you please name where are you setting this password? I have tested with PPP/Secrets and User/Password. Have not managed to reproduce.

Some examples (obviously using 6.31):

Simply open on winbox 2.2.18 interface/wireless/wlan1 panel and change somethihng, nv2 password are replaced with literallly "***********".

Launch bandwidth test (on one device with password) it start,
stop the test and click start again, instead of real password the winbox sent literally "******" as password, because on memory the password are replaced with "*"

Open one wireless security profile, if you uncheck hide password, the passwords are still to "********",
if you close and re-open the panel the password are show.
If you suibmit the panel with "hide password" checked, the passwords are replaced with real "*"

SAme thing with EoIP shared password

Fri Aug 21, 2015 9:46 am

rextended - Thank you! Key here was - winbox 2.2.18

With 3.0rc12 it works fine. I will forward this to our developers.

rextended · Fri Aug 21, 2015 10:13 am

rextended - Thank you! Key here was - winbox 2.2.18

With 3.0rc12 it works fine. I will forward this to our developers.

Thanks, is surely related to this change:
*) winbox - also hide passphrase in CAPsMAN with "Hide Password"

rextended · Fri Aug 21, 2015 10:17 am

minor BUG: (aready know from 6.28 and never fixed):

if you export eoip tunnels with "keepalive=10,10", instead are exported as "!keepalive"

methou · Fri Aug 21, 2015 4:57 pm

Just after 6.31:

Aug/17/2015 14:31:58 system,error,critical router was rebooted without proper shutdown by watchdog timer
Aug/17/2015 15:34:02 system,error,critical router was rebooted without proper shutdown
Aug/18/2015 14:30:29 system,error,critical router was rebooted without proper shutdown
Aug/19/2015 00:24:53 system,info,critical Firmware upgraded successfully, please reboot for changes to take effect!
Aug/19/2015 00:25:17 system,error,critical router was rebooted without proper shutdown
Aug/19/2015 04:51:34 system,error,critical router was rebooted without proper shutdown
Aug/19/2015 04:51:33 system,error,critical router was rebooted without proper shutdown
Aug/19/2015 06:08:36 system,error,critical router was rebooted without proper shutdown
Aug/20/2015 02:13:27 system,error,critical router was rebooted without proper shutdown
Aug/20/2015 02:13:27 system,error,critical router was rebooted without proper shutdown
Aug/20/2015 09:41:00 system,error,critical router was rebooted without proper shutdown
Aug/20/2015 09:59:30 system,error,critical router was rebooted without proper shutdown
Aug/20/2015 09:59:30 system,error,critical router was rebooted without proper shutdown
Aug/20/2015 23:30:26 system,error,critical router was rebooted without proper shutdown by watchdog timer
Aug/21/2015 01:32:33 system,error,critical router was rebooted without proper shutdown by watchdog timer
Aug/21/2015 04:34:44 system,error,critical router was rebooted without proper shutdown by watchdog timer
Aug/21/2015 05:38:25 system,error,critical router was rebooted without proper shutdown by watchdog timer
Aug/21/2015 05:38:26 system,error,critical router was rebooted without proper shutdown by watchdog timer
Aug/21/2015 05:38:25 system,error,critical router was rebooted without proper shutdown by watchdog timer
Aug/21/2015 05:38:26 system,error,critical router was rebooted without proper shutdown by watchdog timer
Aug/21/2015 05:38:25 system,error,critical router was rebooted without proper shutdown by watchdog timer
Aug/21/2015 07:51:27 system,error,critical router was rebooted without proper shutdown by watchdog timer
Aug/21/2015 11:53:43 system,error,critical router was rebooted without proper shutdown by watchdog timer

For the first few times, I managed to disable the watchdog, turns out all clients lost their IP address after lease expired and new client won't get an IPv4 address, v6 works fine, everything other than that does look good to me.

WirelessRudy · Fri Aug 21, 2015 5:12 pm

WIRELESSRUDY http://forum.mikrotik.com/viewtopic.php ... .1#p490862

OK, I've seen the graph.

Resuming now what happens:

- 6.31 is brought out: ===> reboot issue evolves === > that needs a fix.
- MT suggest to use a 6.32rc1 ...... that in my opinion and according the scheme is a release candidate for new to become current version after testing. New features introduced and on the way some bugs might have been fixed in the process.... BUT, as we all know, it might come with yet again with new bugs... or the old ones still not solved.... so why should I test this one. We are not playing roulette here....

I would imagine that if 6.31 has a bug and a fix is made for that bug that has to be tested you would issue a 6.31.1rc and if it tests fine, you have a new 6.31.1 that is a bug fix only....

If you guys now name a new release with a new version number it therefore implicates their will be no more new bugfix only versions of 6.31. But that is actually what we want. We want a current version that came with a bug fixed. Not a new version that also might have the bug fixed but that also might bring new headaches. You guys are too confident in issuing new version. How many bug fixed have we needed over the years?

So instead of trying this new 6.32rc1 I have to revert back to the last current version that has some bugs fixed and had no more bugs for me, at least not the 'reboot' one...... that is 6.30.2

A serious user having problems won't jeopardize his network again by putting a new version 'rc' on a router in the hope it might have fixed his problem and at the same time opening up for new ones.....

Sorry, but that is how it just works in my opinion.

WirelessRudy · Fri Aug 21, 2015 5:35 pm

This looks to be quite dangerous version. It killed one of my testing devices. I still wonder why all the people massively install every new version without any testing in the laboratory before.

Well, let me explain what can happen in bulk upgrades, or manually working as a robot late at night (not to disturb you clients...)

All networks was running 6.27 and over time we tried some 6.28's, some 6.29's and some 6.30's. Every Time because when you do the auto upgrade sequence (or when you issue een massive script order) the software picks the latest 'current' version.

So what happened with us for instance; We were running now some tenths of units (mainly CPE's) on 6.30.2's to our satisfaction and reading the logs it seems to be pretty stable.
So we started to manually (like a robot a 2 am at night) upgrade whole networks..... only to realize after half an hour the new version was now suddenly 6.31! What? Shit, didn't want that. I'd presumed it still was 6.30.2.

So now suddenly I had some 30 units running on 6.31 but stopped upgrading the rest of the network. Let us wait what happens..... Unfortunately the next morning I found that one rb750UP started with rebooting itself after a couple of hours....
Yes, a unit in production, and yes how stupid. But actually because MT is too confident and therefore too fast by putting the new 'current' version as the proposed new to be used upgrade version.....

We are all humans you know. And people can be triggered to make unwanted mistakes just be following wrong procedures or showing too much confidence in product not ready for real world yet.
Hence we have some profession called "ergonomy" which is the study and development of the knowledge on how to design products (that includes software too!) that can not only do a certain job, but that also avoids us stupid humans to make stupid mistakes....

Hence some of us at times put new softwares on production units while we actually don't even wanted it.....
(Off course, if you did because you just wanted the latest, that even more st......)

WirelessRudy · Fri Aug 21, 2015 5:40 pm

In fact, 6.31 is still given as the 'current' version so automated scripts or robots will still load this. But as we have learned now you'd better not! I think the proposed upgrade version should stick at 6.30.2.

And showing the latest current version number with the fix number you also know inmediately you have the latest version included fixed for found issues.....

6.31 should still be a rc

Fri Aug 21, 2015 5:49 pm

Yes, yes but...
The fix must be prepared and tested .... let it be implemented in 6.32rc...if it helps then it should be backported to 6.31.1 or any latest version as it is quite easy to see differences in code and then find the culprit.
They could preapre 6.31.1, let you test it as bug-fixed "stable" but are you sure that they other errors were not be introduced ? I am not.
If you test current development branch for fix then I think you are ready to immediately revert to the latest most stable version despite what version you are testing.

I am not Advocatus Diaboli...it is just my 0.02$

WirelessRudy · Fri Aug 21, 2015 5:55 pm

Also, please, do not fool us that 6.32rcXX is something that would be converted in 6.30.X. Please, keep the versioning, so, if you maintain 6.30 for bugfixes, make the 6.30rcXX and then the correct 6.30.X but keep it distinct from 6.31rcXX and from 6.32rcXX. Do not mix the bugfixing with new features release, so we will never saw sentence like "This will not be corrected in 6.30.X but in 6.31..." Everything should be corrected in the version actually maintained for bugfixes and not delayed to the next version. The subsequent version should always adopt the bugfixes from previous version and not to bring the old bugs again. Otherwise your new huge versioning system is useless.

That's just what I mentioned.

Stick with the version numbers if making bug fixes, even if these are still to be tested ("rc") If than after the test the fix shows good results the original version should be used with the extra figure to show its a bug fix.

If development in the meantime was producing a new version with new features that's fine. But let us network user first start using a version that has some bugs fixed before we decide any new version might be bringing us new features.....

Please be very conservative in producing a new version. 6.31 is obviously not that yet......

WirelessRudy · Fri Aug 21, 2015 5:59 pm

Well, you know what? It's friday afternoon late and probably the guys in Latvia are on their weekend beers already. They probably are not completely used to their new version numbering too. So cheers to you guys, after all its been a big improvement to the past anyway! Some fine tuning and we might enter walhalla! Chears!

InoX · Sat Aug 22, 2015 1:33 am

RB750UP is not accesible or visible in Neighbours since the upgrade. Ruter is responding to ping and is working.

visalink · Sat Aug 22, 2015 1:47 am

RJ45 SFP 10/100 / 1000M copper module in RB922 with RouterOS 6.32RC3 does not work when you plug the network cable into the SFP module, you lose the connection to the Ether1.

Stay tuned...

popovdv · Sat Aug 22, 2015 8:48 am

confirm that 6.31:
1. not stable for 750UP
2. high cpu (100%) load on main CapsMan router when Caps start upgrading

So not stable release 6.31

abis · Sat Aug 22, 2015 1:15 pm

rextended - Can you please name where are you setting this password? I have tested with PPP/Secrets and User/Password. Have not managed to reproduce.
Some examples (obviously using 6.31):

Simply open on winbox 2.2.18 interface/wireless/wlan1 panel and change somethihng, nv2 password are replaced with literallly "***********".

Launch bandwidth test (on one device with password) it start,
stop the test and click start again, instead of real password the winbox sent literally "******" as password, because on memory the password are replaced with "*"

Open one wireless security profile, if you uncheck hide password, the passwords are still to "********",
if you close and re-open the panel the password are show.
If you suibmit the panel with "hide password" checked, the passwords are replaced with real "*"

SAme thing with EoIP shared password

I confirm, happened to me in an IPsec connection type on Peers/Secret (i prefer to use WinBox from The Dude 3.6 but I don't know what version of WinBox it is...)
p.s. Thank you strods for forward!

pukkita · Sat Aug 22, 2015 3:48 pm

Little bug when using System > Packages > Check for Updates on a CCR1009-8G-1S-1S+ w/ 6.30:

14:45:55 script,warning DefConf: Unable to find wireless interface

It prevented the upgraded from being carried on.

Easily fixed by manually uploading the 6.31 npk file, reporting for completeness sake

pbr3000 · Sat Aug 22, 2015 5:09 pm

RB750UP is not accesible or visible in Neighbours since the upgrade. Ruter is responding to ping and is working.

This is exactly the problem I noticed on 6.31rc16 and 6.31 final. Can you make a supout file and send it to Mikrotik support? I could not send because my routers (two, not so important) were in production and I needed go back to 6.30.2.
PS: They stopped working completely after some time and I had to do a manual power cycle.

ngorchilov · Sat Aug 22, 2015 9:42 pm

I have issue with switch port flapping with 6.31 on "model: 750", after downgrade to 6.30.2 all is ok

Same here. RB750 - all switch1 ports (ether2-ether5) are flapping like crazy every 2-3 seconds. Downgraded to 6.30.2 and the issue disappeared. Another try to upgrade to 6.31 - flapping is back. Downgrade agin to 6.30.2 - no flapping.

RB750G as well as few different CCR models are running fine with 6.31.

Tonda · Sun Aug 23, 2015 12:57 am

jarda wrote:
Also, please, do not fool us that 6.32rcXX is something that would be converted in 6.30.X. Please, keep the versioning, so, if you maintain 6.30 for bugfixes, make the 6.30rcXX and then the correct 6.30.X but keep it distinct from 6.31rcXX and from 6.32rcXX. Do not mix the bugfixing with new features release, so we will never saw sentence like "This will not be corrected in 6.30.X but in 6.31..." Everything should be corrected in the version actually maintained for bugfixes and not delayed to the next version. The subsequent version should always adopt the bugfixes from previous version and not to bring the old bugs again. Otherwise your new huge versioning system is useless.

Hm, I am afraid that programmers from Mikrotik will be able to maintain only one working version - this is the way they are used to do for several years. They tried to create new versioning system, but look - there are much more "new" and "rc" versions and only two bugfix versions for 6.30. And look to changelog, how many bugs are solved in versions > 6.30 which are currently still not implemented to 6.30.X. I impatiently wait for bugfixing version 6.30.X which will contain correction of at least CRS port flapping problem (I think I am not alone, because this bug is f*****g annoying). I would like to update my CRS with bugfix version, but currently I do not have any other option than install screwed version 6.31 or some 6.32 "rc" version. So from my point of view (with several years of programming exprerience) they are not able to maintain several versions, they will release new versions in the same way they did it for years. From my point of view this is not good info regarding RouterOS software because they promised something they are not able to achieve (at least in this moment from version 6.30).

Somikov · Sun Aug 23, 2015 1:43 pm

Routerboard RB411
Can not select led for control:

swenske · Sun Aug 23, 2015 2:27 pm

Hello,

I can't download the disk image

http://download2.mikrotik.com/routeros/ ... e-6.31.zip

disk-install.PNG

msdt · Sun Aug 23, 2015 3:51 pm

RB750:
Switch DES-1005P on any port mikrotik.
Auto Negatiation is ON.
Link at intervals of 2 seconds up...down...up....down...
Downgrade to 6.30.2 solved a problem,

RB750, same problem, ports 2 - 5 flaps permanently. 6.30.2 works.

FASINFO · Sun Aug 23, 2015 5:58 pm

Board RB912UAG-5HPnD: There was a problem with the disappearance of remote access to multipoint, after a firmware update to version 6.31: some time lost access via Telnet WnBox and, at the same address multipoint pinged and it generally works. It helps only power off. After some time, access is lost again. After downgrade to 6.30.2 - no problem.

InoX · Sun Aug 23, 2015 6:45 pm

RB750UP is not accesible or visible in Neighbours since the upgrade. Ruter is responding to ping and is working.
This is exactly the problem I noticed on 6.31rc16 and 6.31 final. Can you make a supout file and send it to Mikrotik support? I could not send because my routers (two, not so important) were in production and I needed go back to 6.30.2.
PS: They stopped working completely after some time and I had to do a manual power cycle.

If I can't access the router, I can't make supout either

gtj · Mon Aug 24, 2015 12:22 am

It looks like although getting the script table via snmp is fixed, you can only get and set(run) the one with the highest index...

[root@net-dfw1 snmp]# snmptable -v2c -Ci -c monitor 172.16.0.50 mtxrScriptTable
SNMP table: MIKROTIK-MIB::mtxrScriptTable

 index mtxrScriptName mtxrScriptRunCmd
    15 monitor_config                0
    16     regmonitor                0
    17        startup                0
    18         syslog                0
    
[root@net-dfw1 snmp]# snmpget -v2c -c command 172.16.0.50 MIKROTIK-MIB::mtxrScriptName.15
MIKROTIK-MIB::mtxrScriptName.15 = No Such Instance currently exists at this OID
    
[root@net-dfw1 snmp]# snmpset -v2c -c command 172.16.0.50 MIKROTIK-MIB::mtxrScriptRunCmd.15 = 1
Error in packet.
Reason: (noSuchName) There is no such variable name in this MIB.
Failed object: MIKROTIK-MIB::mtxrScriptRunCmd.15

[root@net-dfw1 snmp]# snmpset -v2c -c command 172.16.0.50 MIKROTIK-MIB::mtxrScriptRunCmd.16 = 1
Error in packet.
Reason: (noSuchName) There is no such variable name in this MIB.
Failed object: MIKROTIK-MIB::mtxrScriptRunCmd.16

[root@net-dfw1 snmp]# snmpset -v2c -c command 172.16.0.50 MIKROTIK-MIB::mtxrScriptRunCmd.17 = 1
Error in packet.
Reason: (noSuchName) There is no such variable name in this MIB.
Failed object: MIKROTIK-MIB::mtxrScriptRunCmd.17

[root@net-dfw1 snmp]# snmpset -v2c -c command 172.16.0.50 MIKROTIK-MIB::mtxrScriptRunCmd.18 = 1
MIKROTIK-MIB::mtxrScriptRunCmd.18 = INTEGER: 1

At least in 6.30.x you could get/set all the scripts if you knew the index.
Big step back.

dcavni · Mon Aug 24, 2015 12:41 am

Well, i waited a few days before upgrading from 6.30, to 6.31 stupidly thinking, that everything should be ok. Well, that was a bad idea. After an upgrade RB2011 worked until i tried to manualy add adress in adress list. Then it rebooted itself and after reboot PPPoE connection didn't work anymore. Somehow i managed to get it together by reseting all the equipment and downgrade back to 6.30.2. Everything is fine now.

6.31 is not ready yet.

dominet · Mon Aug 24, 2015 1:36 am

By upgrading to version 6.31 on an RB751U-2HnD the WiFi and sometimes the system too hangs or completely freezes. I was unable to create the supout.rif because of the system hang. (Firmware version is 3.24)

I have fallen back to version 6.30.2 and there is now no problem at all.

Please let me know if anybody else has the same problem, or if it's a unique hang/freeze on my system only.

Thanks.
DomiNet.

cybernetus · Mon Aug 24, 2015 5:55 am

By upgrading to version 6.31 on an RB751U-2HnD the WiFi and sometimes the system too hangs or completely freezes. I was unable to create the supout.rif because of the system hang. (Firmware version is 3.24)

I have fallen back to version 6.30.2 and there is now no problem at all.

Please let me know if anybody else has the same problem, or if it's a unique hang/freeze on my system only.

Thanks.
DomiNet.

No i have the same problem on my router. My wifi hang/freeze many times and the system too.

I need to revert back to the 6.30.2 too.

AlexS · Mon Aug 24, 2015 8:47 am

What's new in 6.31 (2015-Aug-14 15:42):

*) chr - added support for VMware SCSI virtual disks
*) chr - added support for VMware vmxnet3 network card

Thank you, does that include the open vmware tools as well ????

Just did an upgrade from 29.1 and found my ssh host keys had changed on the box bug or feature ?

EDIT.
shutdown vm added in vmxnet3 nic, it hasn't shown up in the interfaces list

dominet · Mon Aug 24, 2015 9:20 am

Normis,

Unfortunately version v6.32.RC3 still hangs or completely freezes devices RB751U-2HnD after using for a couple of hours.

For several years all standard releases was never a danger to upgrade to, but with version v6.31 you guys just made it. If I were you, I might consider to call back version v6.31 and to keep v6.30.2 as the latest standard release.

Dominet.

Mon Aug 24, 2015 12:22 pm

There were more such problematic updates in the past. I believed that mikrotik changed the approach but it was just a myth from them. I still would like to know what are the quality assurance procedures in mikrotik company.

crash9877 · Mon Aug 24, 2015 6:25 pm

on my RB2011UiAS-2HnD-IN wireless won´t work anymore. (wrong password), Reboot Problem and OpenVpn problems. Had to revert back to 6.30.2
same goes to RB951G-2HnD. this sucks big time.

looks like 6.3.1 was never really tested.

InoX · Mon Aug 24, 2015 6:30 pm

For several years all standard releases was never a danger to upgrade to, but with version v6.31 you guys just made it. If I were you, I might consider to call back version v6.31 and to keep v6.30.2 as the latest standard release.

Dominet.

for some time, standard releases are unstable, and beta/candidate are stable

docmarius · Mon Aug 24, 2015 8:03 pm

If there would be an 6.30.3 with the PPoE upload speed and the CA certificate import fixed, I would be happy with it...
Nah, forget the CA, just the upload speed fix, backport from 6.31...

dominet · Mon Aug 24, 2015 8:08 pm

Not to mention the Wireless hangings and System freezes on the RB751U-2HnD board.

If there would be an 6.30.3 with the PPoE upload speed and the CA certificate import fixed, I would be happy with it...
Nah, forget the CA, just the upload speed fix, backport from 6.31...

pbr3000 · Mon Aug 24, 2015 10:45 pm

RB750UP is not accesible or visible in Neighbours since the upgrade. Ruter is responding to ping and is working.
This is exactly the problem I noticed on 6.31rc16 and 6.31 final. Can you make a supout file and send it to Mikrotik support? I could not send because my routers (two, not so important) were in production and I needed go back to 6.30.2.
PS: They stopped working completely after some time and I had to do a manual power cycle.
If I can't access the router, I can't make supout either

This is so sad... We are on the MT hands now (or brains). Let's wait.

Boter · Mon Aug 24, 2015 10:45 pm

updated all RB's except SXT

so far so good

MTeeker · Tue Aug 25, 2015 12:47 am

Hi Normis, (not sure if you are the recipient of this)

Issue: OpenVPN is broken in 6.31 on my RB2011UiAS-2HnD (mipsbe)

Attempts to connect to router (v6.31) via openVPN is no longer possible. Same result from various devices, PC, Android and smartphone. Connection via OpenVPN works previously on 6.30.1.

I want to report this issue of OpenVPN on 6.31 as similarly experienced by mmabob (posted on Fri. Aug. 14, 8:44pm) and elgrandiegote (Sat. Aug. 15, 5:53am).

The message from my router:
ovpn,debug,error,,,,,,,,,l2tp,info,,debug,,,critical,,,,,,,,,,,,,wa
rning duplicate packet, dropping

I notice an item on change log of 6.31 and wonders if it's somehow related to the broken ovpn:
*) ovpn - better multicore support for interface initialization/authentication/creation.

I am more than happy to do anything else, if needed, in respect of this issue.

Looking forward to hearing from you.

Regards

Tue Aug 25, 2015 2:05 am

All errors should be officially reported to support via email. This is not support forum but user experience forum.

MTeeker · Tue Aug 25, 2015 3:22 am

Thanks for the hint.

Have just reported this issue of OpenVPN broken in 6.31 to MikroTik.

On a related issue. If Fast Track allows traffic to bypass firewall, does this mean that QoS settings such as for VoIP no longer works. Just wondering.

Regards

Tue Aug 25, 2015 9:20 am

Yes, it does mean.

Tue Aug 25, 2015 1:29 pm

We have released 6.32rc5 version. This version include these changes:
*) trafflow - added support for IPv6 packets;
*) switch - fixed port flapping on switch ports of RB750, RB750UP, RB751U-2HnD and RB951-2N (introduced in 6.31)
*) ipsec - added compatibility option skip-peer-id-check;
*) flash - fix kernel failure (exposed by 6.31);
*) bridge firewall - add ipv6 src/dst addr, ip protocol, src/dst port matching to bridge firewall;
*) RB911/RB912 - fix SPI bus lock after fast led blink;
*) ipsec - fix potential memory leak;
*) bridge firewall - vlan matchers support service tag - 0x88a8
*) ippool6 - try to acquire the same prefix if info matches recently freed;
*) crs switch - allow to unset port learn-limit, new default is unset to allow more than 1023 hosts per port;
*) x86 - fixed 32bit multi-cpu kernel support;
*) chr - add hotspot,btest,traffgen support;
*) revised change that caused reboot by watchdog problems introduced in v6.31;
*) ipsec - use local-address for phase 1 matching and initiation;
*) ipsec - fix transport mode ph2 ID ports when policy selects specific ip protocol on initiator;
*) certificates -fixed bug where crl stopped working after a while;
*) ip accounting - fixed kernel crash;
*) snmp - fix system scripts get;
*) hotspot - ignore PoD remote requests if no hs configured;
*) hotspot - fix kernel failure when www plugin aborts on broken html source;
*) torch - add invert filter for src/dst/src6/dst6 addresses ;
*) bonding - add min_links property for 802.3ad mode;
*) snmp - get vlan speed from master interface.

For those of you who have issues with key-exchange-timeout (wireless) - please check if your nv2 key has not been changed to "*****". If you use Winbox2, then with latest version security keys might be replaced with "****".

saaremaa · Tue Aug 25, 2015 3:25 pm

Any news about PPPoE COA implementation?

visalink · Tue Aug 25, 2015 11:01 pm

RJ45 SFP 10/100 / 1000M copper module in RB922 with RouterOS 6.32RC3 does not work when you plug the network cable into the SFP module, you lose the connection to the Ether1.

Stay tuned...

This problem no longer exists in version 6.32RC5

misant · Wed Aug 26, 2015 10:00 am

6.31 version works fine on my test lab equipment RB750G after firmware upgrade to 3.24 .
Previous version of firmware was 2.26 and I was getting port flapping, kernel caused reboots and other unwanted issues.

It looks like Mikrotik staff tests only with fresh firmware, btw everyone never updates firmware without reason. If there are possible issues, you need to inform user with LARGE RED LETTERS that old firmware is not recommended with new release.

Can anyone confirm that 6.31 is ok with last firmware on their test environment.

Anyway I am still unsure if it is safe to install it in production environment.