v5rc6 released

Thu Dec 23, 2010 2:11 pm

v5rc6 (aka v5ip6) is released just in time for holidays.

What's new in 5.0rc6 (2010-Dec-23 13:33):

*) ntp server now requires a broadcast address for broadcast mode to work
it should be specified under broadcast-addresses (X.X.X.255),
255.255.255.255 is not going to work;
*) ssh - fixed showing forwarding settings,
check your setting after upgrade!
*) allow to specify multiple networks (including IPv6)
in user's allowed-from list;
*) added support for Safe Mode to WinBox;
*) allow to specify non acceptable groups for RADIUS console logins;
*) handle correctly IPv6 address on bridged or switched interfaces;
*) added IPv6 connection list;
*) added IPV6 support to graphing;
*) added IPv6 support to SNTP client;
*) added IPv6 support to NTP client & server;
*) added IPv6 support to RADIUS client;
*) added IPv6 support to snmp;
*) added IPv6 support to tftp server, also changed logging levels;
*) added IPv6 support to ssh client;
*) added IPv6 support to watchdog;
*) added IPv6 support for /tool e-mail;
*) added IPv6 support to log remote target;
*) added IPv6 support to API;
*) added IPv6 support to winbox;
*) added IPv6 support to ftpd;
*) added IPv6 support to netwatch;
*) added IPv6 support for /queue simple;
*) added IPv6 support for /tool bandwidth-test;
*) added IPv6 support for /tool torch;
*) added IPv6 support for /interface vrrp (version 3);
*) added IP/IPv6 over ipv6 tunnel interface (/interface ipipv6);
*) added Ethernet over IPv6 tunnel interface (/interface eoipv6);
*) snmp - added ipv6AddrTable from IPV6-MIB;
*) wireless nv2 - fixed station-wds mode multicast problem;
*) wireless nv2 - fixed occasional encrypted link stalls;

http://www.mikrotik.com/download/router ... c6.torrent
http://www.mikrotik.com/download.html

As usual, torrent will work sooner, as mirrors are still being updated (wait for a few hours for the HTTP mirrors to be up to date).

Eising · Thu Dec 23, 2010 4:27 pm

What a nice Christmas surprise! Good job on the IPv6 fixes!

fewi · Thu Dec 23, 2010 4:28 pm

What a nice Christmas surprise! Good job on the IPv6 fixes!

What he said.

FIPTech · Thu Dec 23, 2010 5:19 pm

Good news. Thanks Mikrotik.

DSCP setting for IPv6 IPinIP tunnels is very good. We should now be able to use policy routing with tunnels using DSCP as a matcher. I hope to see this for IPv4 as well.

For information :

IPv4 exhaustion counter has been advanced to 11 february 2011.

There is now only seven /8 blocks available at IANA. APNIC has aanounced a few days ago that they will deliver two /8 blocks quite fastly.

After those 2 /8 blocks, the 5 remaining blocks will be allocated to 5 RIR regions at the same time.

IPv4 end could be seen before the computed date.

Asia will eat their IPv4 ressources very fastly. And we'll need about one year before final RIRs exhaustion.

Good time to begin testing, clients and providers deployment.

Hopefully we have Mikrotik Router OS :=)

Trezona · Thu Dec 23, 2010 6:13 pm

There seems to be a problem with v5rc6!

When i try and mac ping, winbox closes!

Thu Dec 23, 2010 6:14 pm

There seems to be a problem with v5rc6!

When i try and mac ping, winbox closes!

report to support with a supout.rif file please

rpingar · Thu Dec 23, 2010 6:23 pm

There seems to be a problem with v5rc6!

When i try and mac ping, winbox closes!

I confirm this bug too!
winbox crashes when macping something.
To reproduce it:
1. go on ip neighbors;
2. on the list of nighbors, go on the selected item right click on the mouse and choose ping or macping (it is the same)

then winbox will crash.

regards

Ros

FIPTech · Thu Dec 23, 2010 6:29 pm

Seems like Winbox 2.2.16 does not allow IPv6 addresses.

The 5.0 rc6 changelog say :

*) added IPv6 support to winbox;

Is there is new Winbox version available for IPv6 access ?

Thu Dec 23, 2010 6:34 pm

You have to download from the router latest winbox loader v2.2.18

rpingar · Thu Dec 23, 2010 6:41 pm

You have to download from the router latest winbox loader v2.2.18

it crashes also using 2.2.18

regards
Ros

Thu Dec 23, 2010 6:46 pm

What exactly crashes? My post was answer to FIPTech question about winbox IPv6 connectivity. If you are referring to MAC ping winbox crash, then please contact support at mikrotik.com

rpingar · Thu Dec 23, 2010 6:47 pm

What exactly crashes? My post was answer to FIPTech question about winbox IPv6 connectivity. If you are referring to MAC ping winbox crash, then please contact support at mikrotik.com

sorry i tought you were answering my post.

sorry again.

Ros

FIPTech · Thu Dec 23, 2010 6:47 pm

Still no support for linkloop 802.2 protocol.

linkloop 00:0C:42:xx:xx:xx
Link connectivity to LAN station: 00:0C:42:xx:xx:xx (HW addr 0x000C42xxxxxx)
** TIMEOUT (2 seconds)
-- NO RESPONSE --

Would be usefull for testing level 2 connectivity with switches.

Works even if ARP is disabled.

http://sourceforge.net/projects/linkloop/

Thu Dec 23, 2010 7:40 pm

Trezona, MAC-Ping issue is confirmed, we will fix it in the next RouterOS version.

Trezona · Thu Dec 23, 2010 8:41 pm

Trezona, MAC-Ping issue is confirmed, we will fix it in the next RouterOS version.

Thanks guys!

hrvoje · Thu Dec 23, 2010 10:21 pm

on x86 on 3 sistems runing 5.06 Simple Queue stop working

after downgrade everithing is ok

Thu Dec 23, 2010 11:02 pm

on x86 on 3 sistems runing 5.06 Simple Queue stop working

after downgrade everithing is ok

what do you mean by "stop working"? what stops working? the router? queues? something else?

please be specific!

FIPTech · Thu Dec 23, 2010 11:07 pm

Winbox IPv6 seems to run OK.

Very good point.

AlexLV · Thu Dec 23, 2010 11:21 pm

I can confirm about simple queues.
As i think it's because in all my simple queues appears 0.0.0.0/0 value in "Dst. Address" field, and i can't remove this, previous this field was unset.
Normally my simple queues works by packet marks and "Dst. Address" in not necessary.
After downgrade to v5.0rc5 queues works fine.

fewi · Thu Dec 23, 2010 11:28 pm

It's probably related to

*) added IPv6 support for /queue simple;

If it can now handle both IPv4 and IPv6 bugs may have been introduced. That said, 0.0.0.0/0 should be equivalent to an unset destination address since that matches all possible IPv4 addresses, and you're presumably not just having issues with IPv6 traffic.
Out of curiosity, does explicitly adding 0.0.0.0/0 as a destination address work on RC5? When you leave it in like that and then upgrade, does it work in RC6 (is there a difference between explicitly defining it and having it appear as a default value)?

hrvoje · Thu Dec 23, 2010 11:30 pm

on x86 on 3 sistems runing 5.06 Simple Queue stop working

after downgrade everithing is ok
what do you mean by "stop working"? what stops working? the router? queues? something else?

please be specific!

Simple Queue stop working
supoutrif alredy sent

torrents went mad

stanioomega · Fri Dec 24, 2010 12:37 am

and again without working UserManger

on x86

roadracer96 · Fri Dec 24, 2010 1:26 am

Trezona, MAC-Ping issue is confirmed, we will fix it in the next RouterOS version.

Its more than that. If I click on a SSTP active connection on a SSTP server, and then ping, it crashes there too....

roadracer96 · Fri Dec 24, 2010 1:41 am

For the queues, it looks like it always wants an address.. If you try to make it blank via CLI, it puts ipv6 ::/0 in automatically.

elnagar_ali · Fri Dec 24, 2010 1:56 am

hello
after i upgrade to rc 6
my active users have no speed limit
Queue stop working
all user have the full speed with no limit
this problem with user manager users only
but speed limt working fine with pppoe user

Beccara · Fri Dec 24, 2010 2:43 am

*faints* Nice work, Let the bughunt begin!

roadracer96 · Fri Dec 24, 2010 3:07 am

Anyone having anything "weird" happen with IPv6 RA/discovery?

I have an RB1100 with a RB433AH access point behind it. The 1100 bridge interface is enabled for discovery/managed config/other, the RB433 IPv6 address/interface is just mac discovery. But my laptop picks up the 433 as the default gateway, so i need a default route on the 433 for my laptop to get access across the 1100... on 5.0rc5, I didnt need a default gateway on the AP and it picked up RA messages from the 1100..

I dunno what is supposed to be right. Seems like the AP should be a bridge and the 1100 would be the default gateway for my laptop, not the 433...

EDIT: If I disable the IPV6 package on the 433, The 1100 comes through as the default gateway. Even if I remove any global IPV6 address from the access point and remove anything about discovery, it still gets advertised..

Im thinking there is a quirk with advertisement in 5.0rc6. The config worked in 5.0rc5..

tanjil · Fri Dec 24, 2010 4:03 am

in version r5.6 (latest )... queuee limit of ppp useres are not working...

hw to fix?

A

roadracer96 · Fri Dec 24, 2010 4:16 am

Cant add/edit ipv6 routes in winbox, only from cli. Gets an error.

NetworkPro · Fri Dec 24, 2010 5:38 am

I can confirm that the User Manager has not been fixed, the issue is the same like with version 5.0RC5.

- No Store can be activated for User Manager
- User Manager has no idea about any user or anything (as result of the above obviously)

Please fix with lightning speed.
Thank you.

petro25 · Fri Dec 24, 2010 6:12 am

rc5 rc6
rb411 r52hn nstrime 56km SNMP

no route
bridg wds
Problem 100 % CPU
http://78.40.84.226:12111/graphs/

saaremaa · Fri Dec 24, 2010 7:08 am

After update from 5.0rc5 → 5.0rc6 we got a problem with:
1. 5.0rc6 → CPU Load at 100% on SSH Logout not fixed!
2. Simple Queue not working.

tanjil · Fri Dec 24, 2010 10:04 am

saaremaa which version u r using and happy ?

saaremaa · Fri Dec 24, 2010 10:17 am

saaremaa which version u r using and happy ?

5.0rc5, but due to an error I do not use SSH.

Beccara · Fri Dec 24, 2010 10:37 am

I hope everyone is submitting rif files to support. Be a little fair as this is a RC still and has *ALOT* of v6 support added to things which is going to cause problems, Log support tickets and let them work on them. This RC has alot of things changed to alot of key bits (and old code)

stanioomega · Fri Dec 24, 2010 1:03 pm

I can confirm that the User Manager has not been fixed, the issue is the same like with version 5.0RC5.

- No Store can be activated for User Manager
- User Manager has no idea about any user or anything (as result of the above obviously)

Please fix with lightning speed.
Thank you.

confirmed ...

ThunderBRZ · Fri Dec 24, 2010 1:42 pm

I have upgrade rc5 -> rc6 and

and have a hotspot problem !!!

my users have this msg in browser :

"Error 503: unknown method"

i try reset my html and no works.

any have a same problem???

[]´s

FIPTech · Fri Dec 24, 2010 2:55 pm

Seems like TFTP is broken.

See more details here

http://forum.mikrotik.com/viewtopic.php?f=1&t=47763

tanjil · Fri Dec 24, 2010 3:09 pm

so many error on 5.0rc6

and also on rc5...

Alfabi · Fri Dec 24, 2010 3:23 pm

BUG - PPPoE server queues not working too.... all customers without limits....

w0lt · Fri Dec 24, 2010 3:42 pm

You have to download from the router latest winbox loader v2.2.18

When I download Winbox, it still says version 2.2.16. Is there a different location to get 2.2.18?

NetworkPro · Fri Dec 24, 2010 3:48 pm

http://demo2.mt.lv/winbox/winbox.exe ~ latest test version

FIPTech · Fri Dec 24, 2010 5:11 pm

w0lt, where did you get Winbox ?

if you downloaded it from the router web GUI, it should give you 2.2.18. If not then i think you don't have RC6 on the router.

roadracer96 · Fri Dec 24, 2010 5:35 pm

I can confirm the hotspot error 503.

stanioomega · Fri Dec 24, 2010 5:45 pm

I can confirm the hotspot error 503.

same here on x86

mindragon · Fri Dec 24, 2010 6:44 pm

Any ideas when DHCPv6 will be implemented?

Fri Dec 24, 2010 7:20 pm

Any ideas when DHCPv6 will be implemented?

It is in TODO list and should be ready until final v5 is released. You can easily start to deploy IPv6 even without DHCPv6, use stateless address autoconfiguration.

Athan · Fri Dec 24, 2010 10:11 pm

Just to add something positive to this thread, first time during v5 testing I don't have any wireless "sent deauth" issues with iPhones and iPads.
I don't know what you folks have changed in rc6 wireless code but seems it helped in overcoming that nasty issue. I hope it won't come back in future releases.

tudorik · Fri Dec 24, 2010 10:55 pm

Hello,

I don't know if it's a bug or I need to configure something, but when I try to execute a script over ssh that is ~850 lines long in v5rc6, I get the message "Write failed: broken pipe". The script is with customer firewall and queue rules. The same script had no problem on v4.13 and if I shorten it to just a few customers (3 out of all 70) it works fine. (didn't try to see how many would work)
Any ideeas?

Thank you.

EDIT: I've downgraded from v5rc6 to v4.16 and now I get this error when ssh-ing into the system: " Protocol major versions differ: 2 vs. 1 "/ I also noticed that when the system first boots after the downgrade, it only generates an RSA key. After a clean v4.16 install, everything works fine.

andrijco · Sat Dec 25, 2010 8:47 am

hello how to propertly downgrade to version 5.0rc5 if i installed 5.0rc6 from CD ?
i try to add .NPk files from older version and issue command /system package downgrade , but after system rebooted is still rc6 version and simple queues broken, i also try to disable ipv6 package and after Shutingdown services system hangs.

Sat Dec 25, 2010 10:52 am

I can confirm that there are _no_ changes to user manager for the last couple of versions (except where indicated in the changelog). we will start fixing it in one of the upcoming versions, sorry for the delays.

birendersinghbudhwar · Sat Dec 25, 2010 10:59 am

Hi normis,

Any fix for hotspot 503 error in v5RC6 for now i downgraded to v5RC5 but i want to upgrade to v5RC6 but all hotspots customer are getting 5.3 and 400 errors no auto redirects to mikrotik hotspot login page if manually entered the IP address then login page is coming.

Birender

mangust · Sat Dec 25, 2010 9:24 pm

Any ideas when DHCPv6 will be implemented?
It is in TODO list and should be ready until final v5 is released. You can easily start to deploy IPv6 even without DHCPv6, use stateless address autoconfiguration.

To MT:
First of all I would like to say thanks for your job.
You doing very useful job for community. However,
The question is – do you have any schedule for final release?
There were so many changes since 5beta1 (03.2009). When you are going to „freeze” any add-ons into RC?
I can agree that ipv6 is the feature.
But seriously, I don’t need any surprise regarding changes in simple Q or Access lists in case if I chose to upgrade my ”IPV4 only system „ to last RC.
From one side, you have asking community to test RC in real life system; from other side each release has add-on that could break some existing configuration.

I know that IPV6 has so many “stuff” that could be VERY useful for routers.
But I just want to believe that you are not going to add any IPv6 functionality into stable V.5 after it will be released.
Really, could you please just stop with changes, and provide to community stable release with
a)New kernel drivers
b)RPS support
c)N2 support
d)SSTP support
e)GRE support
Do you think these changes are not enough for new version?

chronos · Sat Dec 25, 2010 9:40 pm

What is purpose of release candidate if new features are introduced? This is usually called alpha or trunk branch. RC version should be branch created from trunk where only fixes are applied. And after some time if no more problems are discovered then it could be called as stable branch. So mikrotik should supply trunk, rc and stable versions. New stable version should be released for example in half year period.

In current status it seems to me that final release is far away from now. May be several months because real feature freeze stage not started yet. Even that some particular new features are tested and ready for use.

hrvoje · Sat Dec 25, 2010 10:06 pm

there is stable version v4, and on all mission critical routers this is the way to go, its great that they devolp such a chep and good peace of software and ill will always test it on some routers that i can fixit easily if something go wrong

mangust · Sat Dec 25, 2010 11:01 pm

there is stable version v4, and on all mission critical routers this is the way to go,

New Intel 1G card, SSTP, RPS, GRE?

tanjil · Sun Dec 26, 2010 12:04 am

can i use user manager on mikrotik 5.0rc5 ..? i cant configure because user manager is not working there..
anybody can help me pls?

changeip · Sun Dec 26, 2010 4:43 am

v4 is still premature ... it still has life left hopefully.

v5 started out beta, no one used it because beta is super beta in mikrotik-land... so they made it RC so at least some people would use it, which is why we are finally flushing the bugs out. just my guess : )

JustinTx · Sun Dec 26, 2010 8:22 am

x86 has been running 5.0beta4 for months without a problem.

Upgraded to v5rc6.

[admin@MikroTik] /tool user-manager> database clear
Dangerous! Reset whole user-manager database (except logs)? [y/N]:
y
Resetting user-manager database
Failed to reset user-manager database
error - contact MikroTik support and send a supout file (2)
[admin@MikroTik] /tool user-manager>

I can't believe something so simple or obvious could have been missed during developer testing, so I'm going to *first* assume the problem is at *my* end before I scream it's a bug. My original license was for 3.x, and I'm going to reinstall 3.x and do an upgrade to v5rc6 and see if I can duplicate the problem.

NetworkPro · Sun Dec 26, 2010 8:57 am

Save your efforts. The User Manager RC5 and RC6 is not working at all. Don't you read the posts here?

andrijco · Sun Dec 26, 2010 1:43 pm

who can test if on 5.0rc6 works command like this:

/ip route print count-only where dst-address "0.0.0.0/0"

on 4.6 this command gives result in few seconds
on 5.0rc5 result is infinite processing without any reslut "helps only ^+C"

if this is a bug how to report ?

few of my scripts don't work like

:foreach subnt in=[ /ip route find comment="IX" ] do={
:local bogon;
:set bogon [/ip route get $subnt dst-address];
/ip firewall address-list add list=IX address=$bogon; };

I would appreciate any response.

NetworkPro · Sun Dec 26, 2010 1:58 pm

Hello andrijco,

Your test results are greatly appreciated.

Does the same problem happen with version 5.0 RC6?

To report this, you could simply send an e-mail to support@mikrotik.com telling them the same thing.
Usually you should include a supout file from your router that has the problem to help identify exact conditions.

Thank you.

andrijco · Sun Dec 26, 2010 2:49 pm

i don't know if this "bug" present in 5.0rc6
i recently downgrade version to 5.0rc5
simple queue functionality it's little important to me
so i ask the community =)
please test this thing in current development version

switch · Sun Dec 26, 2010 3:57 pm

v5rc6 sniffer not work.

sobrado · Sun Dec 26, 2010 4:15 pm

Hi andrijco.

Sorry, no access to real hardware yet -- in a few days I will have two RB/750G so will be able to do better testing. As software developer I understand importance of software testing, even more when working on release candidates. Using the online demo router running 5.0rc6 at mt.lv we have:

[demo@demo2.mt.lv] > /ip route print count-only where dst-address "0.0.0.0/0"
0

Mon Dec 27, 2010 8:03 am

there is stable version v4, and on all mission critical routers this is the way to go,
New Intel 1G card, SSTP, RPS, GRE?

so what do you want, new features or stability? you are contradicting yourself.

pascalw · Mon Dec 27, 2010 8:50 am

After update to v5rc6 my IPv6 Tunnel to Hurricane is not working anymore.
The clients don't get a ipv6 ip anymore and advertise is still on.
I did a downgrade to v5rc5 and the ipv6 tunnel is working again also the clients get the ipv6 ip number.

Mon Dec 27, 2010 10:25 am

Thank you very much for your reports about HotSpot. We are working on the problem.

mangust · Mon Dec 27, 2010 10:29 am

there is stable version v4, and on all mission critical routers this is the way to go,
New Intel 1G card, SSTP, RPS, GRE?
so what do you want, new features or stability? you are contradicting yourself.

Normis, ROS is software that is provided by MT, so you should do, what you want (not me

).
And once again - thank for the ROS.

I'm not contradicting myself.
If you asking me, then I can tell, that NOW I need Stability for V5. Because I need features that V5 has.
I'm with v5 because of New Intel card support. SSTP, RPS, GRE Is also good for me.
If you remember we are waiting for long time until MT fix router reboot in case of Simple Q using.
SSH bug, memory leak in sstp, torch bug. And now we get IPV6 "surprise".
So, for now I can use RC3 - because its looks like most stable for me in current situation.
And Once again - is there any schedule for the final release?

Mon Dec 27, 2010 10:36 am

Full release of v5 will be ready when it's ready.

It's still RC, so you should not expect 100% stability, that's what I meant. For stability use v4.16 which is very stable as far as I know (very small amount of problem reports to support).

Beccara · Mon Dec 27, 2010 10:40 am

IPv6 surprise? There's been dozens of threads about it, Those are features you need right now in your network. Right now we needs these IPv6 features and we dont need your features so MT has a hard time pleasing everyone but right now IPv6 is upon us and in order to stay relevant MT has to have full support for v6.

I think everyone here needs a stable ROS 5.0 final release but those features aren't show stoppers. You can workaround the 1gbit card support by using older cards for now, SSTP is still new and really does nothing new compared with L2TP and PPTP etc. We've had to deploy IPv6 with a number of workarounds and messy hacks to get a stable v6 deployment in place now with ROS 4 I can't believe that people are using RC's in production setups. ROS 5 RC's have only just become stable enough to be powered up on lab gear connected to our network.

There are few features in ROS 5 so far that cant be worked around in ROS 4 for the time being

andrijco · Mon Dec 27, 2010 10:48 am

on demo this is MIPS architechture

[demo@demo2.mt.lv] > ip route print count-only where distance=110
2
[demo@demo2.mt.lv] > ip route print count-only where distance=0
0
value of distance out of range (1..255)
[demo@demo2.mt.lv] >

and on demo it works but second answer is strange

maybe it's because my route list counts over 670000 items

janisk · Mon Dec 27, 2010 11:40 am

sometimes architecture is the key, why bug is manifested on one router and not on the other. I have tried /ip ro pri where dst-address="0.0.0.0/0 (and comment="something") on x86 and it worked, same as for ppc, maybe really, route count is limiting factor here.

Beccara · Mon Dec 27, 2010 12:01 pm

On the side topic of ip route print it would be nice if when you do "where dst-address=blah" you can enter a host IP and have it report back the subnet's that route falls into.

When your dealing with a full route table it's dam hard to find a route when you only know the host, In cisco land you can put the host in and it will spit out routes that host falls into. A number of times now we've had to wait 10min for the full table to load in Winbox to manually find the /24 or /19 route that a /16 netblock has

janisk · Mon Dec 27, 2010 1:21 pm

On the side topic of ip route print it would be nice if when you do "where dst-address=blah" you can enter a host IP and have it report back the subnet's that route falls into.

When your dealing with a full route table it's dam hard to find a route when you only know the host, In cisco land you can put the host in and it will spit out routes that host falls into. A number of times now we've had to wait 10min for the full table to load in Winbox to manually find the /24 or /19 route that a /16 netblock has

you can do in clause

/ip route print where 192.168.86.1 in dst-address

Beccara · Mon Dec 27, 2010 1:23 pm

Thanks! Very handy to know :)

andrijco · Mon Dec 27, 2010 1:51 pm

you can do in clause
Code: Select all
/ip route print where 192.168.86.1 in dst-address

[TOTORO@Core0-Wenet] > /ip route print where 192.168.86.1 in dst-address
Flags: X - disabled, A - active, D - dynamic,
C - connect, S - static, r - rip, b - bgp, o - ospf, m - mme,
B - blackhole, U - unreachable, P - prohibit
# DST-ADDRESS PREF-SRC GATEWAY DISTANCE
0 ADb 0.0.0.0/0 194.44.160.65 20
1 Db 0.0.0.0/0 95.215.156.37 20
action timed out - try again, if error continues contact MikroTik support and
send a supout file (13)
[TOTORO@Core0-Wenet] >
[TOTORO@Core0-Wenet] >
[TOTORO@Core0-Wenet] > /ip route print where "Ukraine" in comment
Flags: X - disabled, A - active, D - dynamic,
C - connect, S - static, r - rip, b - bgp, o - ospf, m - mme,
B - blackhole, U - unreachable, P - prohibit
# DST-ADDRESS PREF-SRC GATEWAY DISTANCE
action timed out - try again, if error continues contact MikroTik support and
send a supout file (13)
[TOTORO@Core0-Wenet] >

janisk · Mon Dec 27, 2010 2:10 pm

andrijco you have full BGP feed there?

andrijco · Mon Dec 27, 2010 2:15 pm

andrijco you have full BGP feed there?

2 full views and one local country view

janisk · Mon Dec 27, 2010 2:35 pm

anyway, create supout.rif and send it to us. Also, check if everything is ok with routing through this router.

Mon Dec 27, 2010 2:50 pm

Thank you very much for the reports about queues at 5.0rc6.

We can confirm two problems with queues right now,

1) When there is no address specified (target-address/dst-address), only IPv6 traffic is queued.
To fix it use any IPv4 address (e.g. 0.0.0.0/0) for target-address/dst-address for /queue simple.
To fix dynamic PPP queues, use /queue tree and dynamic-address-list.

2) /queue simple does not work on x86/miplse at all.

Both problems already fixed and will be included to the next release.

Mon Dec 27, 2010 2:50 pm

Thank you very much for the repot about HotSpot server problems. The problem is confirmed as well.

dolphinik1 · Mon Dec 27, 2010 2:56 pm

Feature request for v5rc7: CoA/PoD for all access types(PPTP/PPPOE/L2TP/OVPN/HotSpot)

Mon Dec 27, 2010 3:11 pm

DM are already there for PPP. There is no plans for CoA at PPP in the next version.

w0lt · Mon Dec 27, 2010 3:22 pm

Thank you very much for the reports about queues at 5.0rc6.

We can confirm two problems with queues right now,

1) When there is no address specified (target-address/dst-address), only IPv6 traffic is queued.
To fix it use any IPv4 address (e.g. 0.0.0.0/0) for target-address/dst-address for /queue simple.
To fix dynamic PPP queues, use /queue tree and dynamic-address-list.

2) /queue simple does not work on x86/miplse at all.

Both problems already fixed and will be included to the next release.

Thanks for this update !!
I was experiencing the problem with "Simple Queues" too.
I'll work around it.

-tp

Beccara · Mon Dec 27, 2010 3:43 pm

CoA's something our dev's have asked me about in the past. It's bad form to have to break a PPPoE to cap/uncap someone and static simple queue's quickly become unmanageable at scale. Can it please be put on the wish list?

sobrado · Mon Dec 27, 2010 3:44 pm

@normis
I agree with you, ROS v4 is the stable release right now while v5 is the experimental one. We cannot expect v5 being production-ready yet. Ok, perhaps a "release candidate" is not the place to implement new features, but as noted by Beccara full IPv6 support is not an option anymore. We are moving to IPv6, and we are moving fast. So we need full support for this protocol right now. In our software project, we lock the tree twice a year to stabilize the code for the may and november releases. At this point, no new features are accepted. But in the case of MikroTik, it would be sad not having IPv6 support until ROS v6 (it is even more important for people that, like me, had adquired two devices with a L4 license). So all these IPv6 related changes are welcome.

A few years ago I got an expensive ($900 USD) printer from a well-known manufacturer. This printer supports IPv6 for ping and embedded web server but not for FTP and telnet that remain IPv4-only services. It does not support SSH either. Since day zero, I sent a few requests to the manufacturer about full IPv6 support and adding an SSH service to this printer and got no answer at all. I am greatly surprised to see how MikroTik not only listens to customer requests and works hard to fix bugs in ROS but also implements features wanted by customers. Good work!

mangust · Mon Dec 27, 2010 7:12 pm

Unknown group demo ...

omidkosari · Mon Dec 27, 2010 8:40 pm

Please implement PPPoE Option 82 resending to radius server . we have a lot of problems here .

MlokeOne · Mon Dec 27, 2010 9:54 pm

Hi, I have also a Hotspot bug, on my testing 750G, MAC, Cookie and HTTP PAP auth. It seems to hangs only by HTTP PAP, because mac and cookies are working fine and logs me on without problem.
Naturally, I wish to all mikrotik developers good luck and many success in new year, especialy with IPv6.

elnagar_ali · Tue Dec 28, 2010 1:47 am

Thank you very much for the reports about queues at 5.0rc6.

We can confirm two problems with queues right now,

1) When there is no address specified (target-address/dst-address), only IPv6 traffic is queued.
To fix it use any IPv4 address (e.g. 0.0.0.0/0) for target-address/dst-address for /queue simple.
To fix dynamic PPP queues, use /queue tree and dynamic-address-list.

2) /queue simple does not work on x86/miplse at all.

Both problems already fixed and will be included to the next release.

now i know why the queues work with pppoe and not working with user manager
because all pppoe users have ip address
and user manager have auto ip
thank u
i will wait rc 7

Tue Dec 28, 2010 8:40 am

mangust, thank you very much for the report.
We are working on the problem.

Tue Dec 28, 2010 10:36 am

Hello,

I don't know if it's a bug or I need to configure something, but when I try to execute a script over ssh that is ~850 lines long in v5rc6, I get the message "Write failed: broken pipe". The script is with customer firewall and queue rules. The same script had no problem on v4.13 and if I shorten it to just a few customers (3 out of all 70) it works fine. (didn't try to see how many would work)
Any ideeas?

Thank you.

EDIT: I've downgraded from v5rc6 to v4.16 and now I get this error when ssh-ing into the system: " Protocol major versions differ: 2 vs. 1 "/ I also noticed that when the system first boots after the downgrade, it only generates an RSA key. After a clean v4.16 install, everything works fine.

can you send your supout.rif file to support? we would like to see your script and how you run it, please give us more details.

tudorik · Tue Dec 28, 2010 11:12 am

I sent a message to support,including a "supout.rif" file. The script I used is just plain "/queue tree add" (and remove), "/queue type add" (and remove), "/ip firewall mangle" rules for customers. The routers are all x86. The 2 I tested the scripts are on Virtualbox 4 with demo licenses, but I don't think that's the problem.
And as I said, it works without problems on v4. The same kind of script, 308 lines long (for another router) works on v4 and v5. (the first was 803 lines).

PS: I'll also send the two scripts to support if it makes a difference.

EDIT: ticket is Ticket#2010122666000091

Tue Dec 28, 2010 12:19 pm

v5rc6 sniffer not work.

tested on all platforms, works fine. make sure you have disk space available, this could cause it to not save anything.

gcsuri · Tue Dec 28, 2010 12:53 pm

Hi,

I've just upgraded from v5beta6 to rc6 and my 532 board not boot:
RouterBOOT booter 2.18

RouterBoard 532

CPU frequency: 399 MHz
Memory size: 32 MB

Press any key within 2 seconds to enter setup..
loading kernel from nand... OK
setting up elf image... OK
jumping to kernel code
Kernel panic - not syncing: No init found. Try passing init= option to kernel. See Linux Documentation/init.txt for guidance.

What to do?

regards,

Tue Dec 28, 2010 12:53 pm

Hi,

I've just upgraded from v5beta6 to rc6 and my 532 board not boot:
RouterBOOT booter 2.18

RouterBoard 532

CPU frequency: 399 MHz
Memory size: 32 MB

Press any key within 2 seconds to enter setup..
loading kernel from nand... OK
setting up elf image... OK
jumping to kernel code
Kernel panic - not syncing: No init found. Try passing init= option to kernel. See Linux Documentation/init.txt for guidance.

What to do?

regards,

how did you upgrade, big package to FTP and reboot? which packages did you upload?

gcsuri · Tue Dec 28, 2010 1:03 pm

yes, via ftp. all packages w/o calea, gps, ipv6, hotspot, ups and user-manager.

Tue Dec 28, 2010 1:48 pm

yes, via ftp. all packages w/o calea, gps, ipv6, hotspot, ups and user-manager.

We have no such problems with RB532. You will have to use Netinstall to restore RouterOS.

roadracer96 · Tue Dec 28, 2010 4:50 pm

SSTP is still new and really does nothing new compared with L2TP and PPTP etc.

Wow. You might wanna watch what you say when comparing L2TP/PPTP to an actual secure VPN solution..

Tue Dec 28, 2010 4:52 pm

We have reproduced the problem with SSTP. We are looking for the fix.

sobrado · Tue Dec 28, 2010 5:02 pm

I've downgraded from v5rc6 to v4.16 and now I get this error when ssh-ing into the system: " Protocol major versions differ: 2 vs. 1 "/ I also noticed that when the system first boots after the downgrade, it only generates an RSA key. After a clean v4.16 install, everything works fine.

ROS 4.16 supports both protocol 1 and protocol 2. Its ssh daemon is based on an ancient OpenSSH release (2.3.0); ROS 5.0rc6 supports protocol 2 only so it seems based on, at least, OpenSSH 5.4. In fact, we decided disabling legacy protocol 1 by default on march's release after a transition period of ten years. Support for protocol 1 has not been removed from OpenSSH's source code, only disabled by default in the configuration files.

Protocol 1 supports RSA keys only, protocol 2 added support for DSA keys.

Edit: see http://www.openssh.org/txt/release-5.4

sobrado · Tue Dec 28, 2010 6:15 pm

I don't know if it's a bug or I need to configure something, but when I try to execute a script over ssh that is ~850 lines long in v5rc6, I get the message "Write failed: broken pipe". The script is with customer firewall and queue rules. The same script had no problem on v4.13 and if I shorten it to just a few customers (3 out of all 70) it works fine. (didn't try to see how many would work)
Any ideeas?

"Write failed: Broken pipe" means that packet_write_poll() cannot write some of the output on a socket because there is no process to read the data on the other side. I would suggest (1) trying another sftp client, and (2) to make sure this is not a ROS-specific problem transferring a file to a BSD or Linux system running OpenSSH 5.4 up to 5.6 (as you noted, this problem does not happen with SSH service on 4.13).

Edit: obviously (2) should be done using the sftp client that is unable to transfer the file to 5.0rc6.

tudorik · Tue Dec 28, 2010 9:41 pm

I'm using OpenSSH v5.6p1 on Gentoo (latest available) and just running commands over ssh to an x86 RouterOS box (about 800 at once...).
I think it's the v5 ssh implementation (as there are no problems in v4) and it will be fixed in rc7 according to this (very prompt

) answer:

Hello,

OK, we will increase the maximum length.

Regards,
Normunds

Beccara · Wed Dec 29, 2010 4:59 am

SSTP is still new and really does nothing new compared with L2TP and PPTP etc.

Wow. You might wanna watch what you say when comparing L2TP/PPTP to an actual secure VPN solution..

My post was about work arounds, SSTP offers nothing other than being simple, if you need secure stable VPN right now work around it, 5.0 is not for production deployment!

roadracer96 · Wed Dec 29, 2010 5:31 pm

SSTP is still new and really does nothing new compared with L2TP and PPTP etc.

Wow. You might wanna watch what you say when comparing L2TP/PPTP to an actual secure VPN solution..
My post was about work arounds, SSTP offers nothing other than being simple, if you need secure stable VPN right now work around it, 5.0 is not for production deployment!

Cant. OpenVPN is for crap (The MT config anyways) in larger VPNs.

FWIW, 5.0rc5 has been serving me pretty well w/ 150+ SSTP connections. I just cant run SSTP on my RB1000 or the memory leak problem is too great. Have to use 2x RB450gs behind the RB1000 to slow the memory leak enough to keep em running. Client-side, 5.0rc5 (For my needs at least) is flawless except for the ssh hangup.

I think MTs biggest problem is lack of focus. They dont focus on one thing, finish it, and move on to the next. They finish something 90%, then start new projects and end up with 12 things at 92% completion all the while, people are trying to use said features.

Thu Dec 30, 2010 7:40 am

SSH stuff is fixed in rc7, try it please

JustinTx · Thu Dec 30, 2010 9:44 am

Save your efforts. The User Manager RC5 and RC6 is not working at all. Don't you read the posts here?

I just noticed a post about that a little before mine, so it's likely I was in the middle of my post and started on something when that got posted.

Who is online