Thanks, got it. My bad, used to see that testing is usually 1 minor version ahead.
Yes, the memory leak appears to be fixed, the resolver now returns the large replies OK, but they still aren't loaded into an address list :-(*) dns - fixed memory leak caused by large DNS replies;
When testing moves from beta## to rc# and after one or more rc# the new stable is released, the stable version gets that major version number (without beta or rc suffix) and the testing remains at rc# for some time until it updates to the next version number and beta##.Thanks, got it. My bad, used to see that testing is usually 1 minor version ahead.
no, when I updated a CCR2004-1G-12S+2XS it turned on auto mode and I lost connectivity.*) sfp28 - changed FEC auto mode to disabled;
What version is your routerboot(firmware) at?FCK MIKROTIK!!!!
Got half a heart attack and half a stroke!!!
The CCR1072 took around 15 minutes and CCR1036 around 10 minutes to come back after the firmware update.
PS: The (SNMP) voltage reading of the CCR1036 is correct again.
Edit: I can confirm, after applying the 6.49 firmware update, every reboot of the CCR1072 will took arounf 15 minutes and of the CCR1036 around 10 minutes.
ROS 6.49What version is your routerboot(firmware) at?
:put [/ip cloud get public-address-ipv6]
That usually is an existing firmware bug, do netinstall.Cant upgrade 2x LHG 5 AC from 6.49beta22 to 6.49 Stable, not enough space, i got nothing on internal storage...
What exactly was fixed here?*) winbox - fixed support for "Delegated-IPv6-Prefix" for PPP services;
How would you be able to connect this model to the device it only has usb and it's not Smart UPS. I thought Mikrotik supports only APC Smart.*) ups - added battery info for APC Back-UPS BX750MI
How would you be able to connect this model to the device it only has usb and it's not Smart UPS. I thought Mikrotik supports only APC Smart.*) ups - added battery info for APC Back-UPS BX750MI
Items needed for the setup:
1) A Mikrotik device with a USB or RS232 port;
2) USB or RS232 cable;
3) APC UPS with a USB port.
*) user - added "expired" user status with suggestion to change password (WinBox v3.29 required);
*) winbox - minimal required version is v3.30;
*) defconf - removed overlapping IPv6 firewall rules;
In the default config there is an address list bad_ipv6 that contained some addresses that were actually subnets of another one. These were removed.What exactly was removed?
*) defconf - removed overlapping IPv6 firewall rules;
Is there info on what has been changed with DFS a bit more precisely?*) wireless - improve regulatory compliance with DFS requirements;
Anyone have IP-IP tunnel + IPSec problem?
After morning update, between my CCR1036 and RB2011 stopped tunneled traffic if tunnel IPSec encrypted.
Nothing other changed on this routers.
This might be a coincidence. But checking our SNMP log, we only see occasional -274° temp values on the two RB4011 updated to 6.49. Not on others still running 6.48.x.Yes, this is unlikely a bug in the new version. We regularly see the RB temperature sensors return absolute zero when they cannot get a proper reading, in any version of RouterOS.
Depending on the path the IPsec traffic takes between routers (e.g. via another NAT router or stateful firewall), IPsec tunnels can sometimes fail when one side is rebooted even without any software change.Same issue with L2TP IPSec tunnels. Error message: "Tunnel was not encrypted"
Is there a solution besides downgrading?
maybe it's related to this?Hi,
Does anyone know what has changed in 6.49 that would stop Rancid from connecting and taking a backup? Initially I thought it was to do with expired user account but have worked through all that and mtlogin can get in but rancid-run fails to pull the config.
Thanks.
*) console - require "write+ftp" permissions for executing script to file;
*) console - require "write+ftp" permissions for printing to file;
The same happens to me.IPv6 address mapping in IP Cloud does not work. Everything worked in RC2.
The same happens to me. RB4011. Temperature -274.Problems with health continues on some devices:
RB1100AHx4: No temperature
mANTBox2: Temperature = -274
Regards.
Stick with long-term, especially on big routers like that that are most likely used for the Core of your network.FCK MIKROTIK!!!!
Got half a heart attack and half a stroke!!!
The CCR1072 took around 15 minutes and CCR1036 around 10 minutes to come back after the firmware update.
PS: The (SNMP) voltage reading of the CCR1036 is correct again.
Edit: I can confirm, after applying the 6.49 firmware update, every reboot of the CCR1072 will took arounf 15 minutes and of the CCR1036 around 10 minutes.
When you have such a router (i.e. not one of the new home routers with 16MB flash), always partition the flash into 2 partitions so you can copy the working partition, then upgrade, and when you have issues like that you can switch to the other partition and have everything back untouched. No need to do downgrades, and when someone posts a workaround here you can switch back and try it.Hello,
I upgraded my CCR1009 yesterday… and lost my internet connection.
My SERCOMM FGS202 SFP Module isn't recognized anymore. Problem is, it's a quite widely used one in France since the main ISP Orange ships it with their Internet box.
Is there a nice tutorial to do that without screwed everything up ?When you have such a router (i.e. not one of the new home routers with 16MB flash), always partition the flash into 2 partitions so you can copy the working partition, then upgrade, and when you have issues like that you can switch to the other partition and have everything back untouched. No need to do downgrades, and when someone posts a workaround here you can switch back and try it.Hello,
I upgraded my CCR1009 yesterday… and lost my internet connection.
My SERCOMM FGS202 SFP Module isn't recognized anymore. Problem is, it's a quite widely used one in France since the main ISP Orange ships it with their Internet box.
Hello @pe1chlWhen you have such a router (i.e. not one of the new home routers with 16MB flash), always partition the flash into 2 partitions so you can copy the working partition, then upgrade, and when you have issues like that you can switch to the other partition and have everything back untouched. No need to do downgrades, and when someone posts a workaround here you can switch back and try it.Hello,
I upgraded my CCR1009 yesterday… and lost my internet connection.
My SERCOMM FGS202 SFP Module isn't recognized anymore. Problem is, it's a quite widely used one in France since the main ISP Orange ships it with their Internet box.
It's pretty straightforward after reading the documentation: https://wiki.mikrotik.com/wiki/Manual:PartitionsIs there a nice tutorial to do that without screwed everything up ?
When you have such a router (i.e. not one of the new home routers with 16MB flash), always partition the flash into 2 partitions so you can copy the working partition, then upgrade, and when you have issues like that you can switch to the other partition and have everything back untouched. No need to do downgrades, and when someone posts a workaround here you can switch back and try it.
Thx :)
/partitions repartition partitions=2
/partitions print
/partitions copy-to xxx
/partitions activate xxx
/partitions save-config-to xxx
/partitions copy-to backup
Perfect, thx :)No when you repartition to 2 partitions, the first one will contain what your entire flash contains now, and the second one will be empty.
Check that the current disk usage is (much) less than half of the total, and you will be safe.
Yes, I spoke to soon. I occasionally see this -274 temperature value from other devices, but it is only for a brief time when the sensor is not working for some reason. This indeed seems to be an additional issue in this release.This might be a coincidence. But checking our SNMP log, we only see occasional -274° temp values on the two RB4011 updated to 6.49. Not on others still running 6.48.x.
My coworker who uses RANCID is having the same issue in this version. I did a web search and found that people have this issue with 7.1rc3 and rc4 as well.Does anyone know what has changed in 6.49 that would stop Rancid from connecting and taking a backup? Initially I thought it was to do with expired user account but have worked through all that and mtlogin can get in but rancid-run fails to pull the config.
It's pretty straightforward after reading the documentation: https://wiki.mikrotik.com/wiki/Manual:Partitions
Is there a nice tutorial to do that without screwed everything up ?
Thx :)
Thanks. Tried 7.1rc4 and found the same too. Still can’t figure out why (rancid account with full permissions etc.) If you figure anything, please let me know. Going back to 6.48 for now.
My coworker who uses RANCID is having the same issue in this version. I did a web search and found that people have this issue with 7.1rc3 and rc4 as well.
Does this refer to "my" problem with the long reboot sequence?The problem with rebooting is connected with Queues. Spent 7 hours with that! :-( Simple rules are broken and causing problems. Tested on CCR1036 r1.
I hope that is on an isolated lab testing network.I do not use-need password so i wanna vanish those reminders.
There is no password expiry time. The expiration is done manually (using a command or script).Is there a way to increase the password expiry time or remove this check completely?
Where can I find more detailed information about these two items?*) wireless - improve regulatory compliance with DFS requirements;
*) wireless - improve signaling of QCA9984 interface capabilities when using 160/80+80MHz channel width;
...If i remember previous version was 6.37
Now i upgraded RB411U from 6.40 to 6.48.4 to 6.49. First upgrade was perfect, second upgrade to 6.49 and RB stop working, when i was upgrade firmware. There is a problem when upgrade only "system-routerboard-firmware"....If i remember previous version was 6.37
That is a "massive" jump between versions, I suspect there might be config problems on your side
Same Question from My side...What exactly was fixed here?*) winbox - fixed support for "Delegated-IPv6-Prefix" for PPP services;
Yes i know that, but for my setup that i cant explain now, i do not want to use passwords. The problem is that i cant use terminal without setting up a password and not so much the annoying popup. I need the old option for that plz so i can work without everytime need to set password. This is bullying......This only applies to a device with no password at all. In the light of recent attacks against MikroTik devises, why not simply set any kind of password? It will stop asking
I confirm , 6.49 and 7.x doesn't work rancid-run.Hi,
Does anyone know what has changed in 6.49 that would stop Rancid from connecting and taking a backup? Initially I thought it was to do with expired user account but have worked through all that and mtlogin can get in but rancid-run fails to pull the config.
Thanks.
I saw the same issue with my 1072 and came here to see. Going to downgrade until the next version I guess...Thanks for doing the testing for us!Did further investigation regarding the long-boot issue.
ROS 6.48.4 + Firmware 6.48.4 -> no issue -> booting takes around 1:30 minutes
ROS 6.49 + Firmware 6.48.4 -> no issue -> booting takes around 1:30 minutes
ROS 6.49 + Firmware 6.49 -> issue -> booting takes around 10 minutes!
I have done now a few down/upgrades, always with the same effect.
The shutdown sequence will durate about 8:15 minutes. The following boot is as fast as normal in about 1:30 minutes.
This issue is reproducible on all our CCR1072 and CCR1036 devices. Not reproducible on our CRS and RB devices.
Still not fixed on my rb951G-2hnd. Have to turn off 'Verify DOH Certificate'Yes, the memory leak appears to be fixed, the resolver now returns the large replies OK, but they still aren't loaded into an address list :-(*) dns - fixed memory leak caused by large DNS replies;
The issue I reported and was fixed was not related to DoH. Maybe there are still issues in that.Still not fixed on my rb951G-2hnd. Have to turn off 'Verify DOH Certificate'
Yes, the memory leak appears to be fixed, the resolver now returns the large replies OK, but they still aren't loaded into an address list :-(
I saw the same issue with my 1072 and came here to see. Going to downgrade until the next version I guess...Thanks for doing the testing for us!
Oh, ok, that is fine.This only applies to a device with no password at all. In the light of recent attacks against MikroTik devises, why not simply set any kind of password? It will stop asking
I'm running GRE+IPSec behind the NAT on both sides with between MikroTIK devices of a different models. IKE1 with MAIN exchange mode. All my 40 tunnels broke overnight after automated installation of long-term 6.48.5. It seems that initially IPsec tunnels establish just fine, but hang on the next Phase1 (1 day in my case) . The only temporary fix for me now is to manually clear all tunnels once a day which is a pain in my arse. I can't figure out what to do and how to reconfigure it =(Anyone have IP-IP tunnel + IPSec problem?
After morning update, between my CCR1036 and RB2011 stopped tunneled traffic if tunnel IPSec encrypted.
Nothing other changed on this routers.
Hi,I upgraded my CCR1009 yesterday… and lost my internet connection.
My SERCOMM FGS202 SFP Module isn't recognized anymore. Problem is, it's a quite widely used one in France since the main ISP Orange ships it with their Internet box.
@Normis, So in a big installation where we install 160+ devices that are configured script bassed from the main after reset and upgrades, it will ask for password. This is supid and require us to do extra steps.This only applies to a device with no password at all. In the light of recent attacks against MikroTik devises, why not simply set any kind of password? It will stop asking
Hi,Hi,I upgraded my CCR1009 yesterday… and lost my internet connection.
My SERCOMM FGS202 SFP Module isn't recognized anymore. Problem is, it's a quite widely used one in France since the main ISP Orange ships it with their Internet box.
This is quite strange. I have experienced this bug for more than a year since v6.47 with the same SFP (SERCOMM FGS202) and my RB3011. This release (v6.49) is the first one that make my SFP work.
On my side I am very happy with this release which has allowed me to upgrade from v6.46.6 since it was the last working version for the SFP.
Antoine
Are you running CHR on a virtualization environment? Or is it a physical router?Upgrade from 6.48.4 to 6.49 => Clients do not get IP address via DHCP over the EoIP tunnel
Can you share the overlapping subnets that were removed?In the default config there is an address list bad_ipv6 that contained some addresses that were actually subnets of another one. These were removed.What exactly was removed?
It's been years and even on RouterOS v7, IPv6 support is lacking, no NAT64, no PCP etc.Same Question from My side...
What exactly was fixed here?
Why Mikrotik Team is not responding to anything related to IPv6.... Are they themself not clear about it???
RB750GR3 on both sites.Are you running CHR on a virtualization environment? Or is it a physical router?Upgrade from 6.48.4 to 6.49 => Clients do not get IP address via DHCP over the EoIP tunnel
I had such issues and it was due to permission settings in the virtualization (allow promiscuous mode, allow MAC spoofing, etc)
noted here viewtopic.php?p=881501#p8815116.49 and 7.1RC4 - there is no protected bootloader setting in the "System-RouterBOARD-Settings" section of Winbox
Every post you make is always nothing but general complaints. no specific problems mentioned.Hey Mikrotik,
When I looked on some threads, you make very basic bugs in every new firmware...Is this normal programming? Because when I buy some crap Asus, TP-Link...it has more stable firmware, than this...
In my view you don't understand, what you programming or what is some function for...Because if you know it, there will be less bugs in new firmware. Maybe will be better do less function, but stable than many features, but buggy and useless..Testing longer and than make publication like new firmware to people. Because it is not beta and some ISP have hard time, for repair it, or the internet wan't function correctly and the people are angry! And why..because ISP want to have newer FW and not so buggy. So make some new jobs for people, programming some AI for learning how to make work some router correctly and the AI will corect some code in your FW. Because the AI will be testing more functions, than human, if you don't have some time to testing it. This is not funny for me, as a costumer. Because when I setup some function and the setup not working, it's hard for me, because I don't know, if I make some mistake, or it's only bugs in your firmware. Thank you for understanding.
nobody can help you except to offer some cheese and crackers with your wine.... if normal consumer brands are ok for you, as you mentioned, then stop complaining and go get something that works for you. Mikrotik devices are not easy, nobody says they are. but they are far more powerful and flexible than any other piece of networking equipment I've ever used...including Openwrt or DDwrtYes, but when I don't need all this functions, so [TPlink, Asus, etc] is ok for me.
Hello,Relax, you are in the wrong topic. RB5009 is v7 only and here is 6.49 topic. ROFL
Out of curiosity, do you any specific specific sfp-rate-select settings for your FGS202 interface ?
sfp-rate-select=high
I just got an answer from support with a fix. Works fine.Just upgraded a bunch of devices from 6.48.4 to 6.49
on my CRS125 I lost temperature from system health ... (and so on snmp ...)
created a support ticket #[SUP-63119] on this
Hello,Out of curiosity, do you any specific specific sfp-rate-select settings for your FGS202 interface ?Code: Select allsfp-rate-select=high
/interface ethernet set sfp1 sfp-rate-select=low
Thank you.SFP rate low/hi is state on a specific SFP slot pin. Actual function depends on specific SFP module model and should be noted in a documentation/datasheet. There is no standardized meaning what low and high mean, that's up to the SFP manufacturer to define.
OK then. PSA:I updated two Mikrotik 2011UiAS models. The first reboot was fine, but neither have come back from the firmware update reboot. They're off-site, so I haven't had a chance to discover what state they're in.
Hello,Thank you.SFP rate low/hi is state on a specific SFP slot pin. Actual function depends on specific SFP module model and should be noted in a documentation/datasheet. There is no standardized meaning what low and high mean, that's up to the SFP manufacturer to define.
So If I don't know, how to setup it, what is better setup? Or default rate is? The bad setup only don't work, or function is ok, but sometimes with troubles..?
Hello, thank you so much, for explanation and for your time :)Hello,
I've currently no idea of what sfp-rate-select does exactly (except that it uses a different pin on the module, whatever it could mean).
I've asked Mikrotik support for precisions.
About the "best" setting to use, my answer would be: if it works, keep current setting. If not, meaning if your SFP isn't recognized and/or unstable, try to change from high (default value) to low.
What I can tell however is that it has no visible impact on performances.
Best,
Jean Baptiste Favre
That sounds great, but we don't have time to program something new at the drop of a hat, with zero warning that a long-running solution will just stop working.It looks like rancid is using a telnet or ssh session to talk to the command interpreter intended for humans.
That is not a wise thing to do. It is better to write a client that uses API. That could take the place of that mtlogin program and extract the information in a much more repeatable way (and make it much easier to parse as well).
I would have hoped the changelog commented on this change AND the documentation saw updates.When this is not done, you are at the mercy of the control character processing and it looks like in this case the Hotlock mode (used to be Ctrl-V, this has changed, but unfortunately MikroTik have not updated the documentation) has been inadvertently enabled.
Even in an export verbose there's no default parameter for hotlock in any conditions, which is far from unreasonable to request if they're going to change the default. But thankfully dumb mode seems to work around the problem. If anyone is stuck with this, modify your cloginrc file to include +cte in the user. This is what works for me:From the sequences above I guess that the login to the router is not using any of the options that are available to keep the console UI as simple as possible.
When logging in from software like this, you should append options to the username, see https://wiki.mikrotik.com/wiki/Manual:C ... in_process
At least you can try to login using admin+cte (or whatever_username+cte) to suppress the cursor control and color stuff.
However that does not disable hotlock mode, unfortunately. And as it is a toggle, you never know if it is enabled or not.
Of course there should have been an option to turn it off or it should be always off in dumb mode.
add method host ssh
add user host admin+cte
add password host password
Not always practical, and even if people have something like that in development they probably need their configs backed up in the interim.But it remains a sub-optimal solution relative to using API, please use that.
OK then. PSA:I updated two Mikrotik 2011UiAS models. The first reboot was fine, but neither have come back from the firmware update reboot. They're off-site, so I haven't had a chance to discover what state they're in.
___DO NOT UNDER ANY CIRCUMSTANCES UPGRADE FIRMWARE ON RB2011 TO 6.49___
___DO NOT UNDER ANY CIRCUMSTANCES UPGRADE FIRMWARE ON RB2011 TO 6.49___
___DO NOT UNDER ANY CIRCUMSTANCES UPGRADE FIRMWARE ON RB2011 TO 6.49___
You will have to get physical with the device if you do this.
> system router print
routerboard: yes
model: 2011UiAS-2HnD
serial-number: <redacted>
firmware-type: ar9344
factory-firmware: 3.33
current-firmware: 6.49
upgrade-firmware: 6.49
Well, roll the dice few more times then. I had 100% "success" ratio on my 2 attempts. Few others above were also "lucky". ;)
OK then. PSA:
___DO NOT UNDER ANY CIRCUMSTANCES UPGRADE FIRMWARE ON RB2011 TO 6.49___
___DO NOT UNDER ANY CIRCUMSTANCES UPGRADE FIRMWARE ON RB2011 TO 6.49___
___DO NOT UNDER ANY CIRCUMSTANCES UPGRADE FIRMWARE ON RB2011 TO 6.49___
You will have to get physical with the device if you do this.
It seemed to work for me...Code: Select all> system router print routerboard: yes model: 2011UiAS-2HnD serial-number: <redacted> firmware-type: ar9344 factory-firmware: 3.33 current-firmware: 6.49 upgrade-firmware: 6.49
You did do a second reboot after upgrading firmware to 6.49, right?In the morning I upgraded 200 sxts from 6.48.4 also fw 6.48.4 to 6.49 with fw and without any problem
For anything business-critical, it's always a good idea whatever equipment or supplier.It may be a good idea to turn off auto-upgrade until this mystery has been solved.
From my observations actually bricking the device requires firmware upgrade. But good catch with check-installation. I'll do some more experiments.Same problem, 130 devices on remote locations to fix with net install...
After upgrade to 6.49 or 6.48.5 do, a check installation (winbox) in system packages. If it says that installation is corrupted, you have a 100% chance to brick device after another reboot.
To avoid this, when installation is corrupted, downgrade it do 6.48.4, after that upgrade again to 6.48.5 or 6.49 and then Check installation will show You, that everything is OK. From now Next reboot are safe.
Upgrading firmware has nothing to do with it. I think that problem is with upgrade path between some old to new (latest) versions.From my observations actually bricking the device requires firmware upgrade. But good catch with check-installation. I'll do some more experiments.Same problem, 130 devices on remote locations to fix with net install...
After upgrade to 6.49 or 6.48.5 do, a check installation (winbox) in system packages. If it says that installation is corrupted, you have a 100% chance to brick device after another reboot.
To avoid this, when installation is corrupted, downgrade it do 6.48.4, after that upgrade again to 6.48.5 or 6.49 and then Check installation will show You, that everything is OK. From now Next reboot are safe.
/routing bgp instance
set default as=12345 router-id=1.2.3.4
add name=bgp2 router-id=2.3.4.5
You did do a second reboot after upgrading firmware to 6.49, right?In the morning I upgraded 200 sxts from 6.48.4 also fw 6.48.4 to 6.49 with fw and without any problem
/interface wireless channels
add band=2ghz-b/g/n frequency=2412 list=1-6-11 name=ch1 width=20
add band=2ghz-b/g/n frequency=2437 list=1-6-11 name=ch6 width=20
add band=2ghz-b/g/n frequency=2462 list=1-6-11 name=ch11 width=20
set [ find default-name=wlan2 ] ....
scan-list=1-6-11 ...
Do you have thousands of static DNS records in your router e.g. for some form of blacklist?after upgrade to 6.49 on CCR1072, boot time went to +10 minutes
On what model of router?I had to downgrade back to 6.48.3 because the 6.49 broke my PPPOE connection
RB4011iGS+5HacQ2HnDOn what model of router?I had to downgrade back to 6.48.3 because the 6.49 broke my PPPOE connection
The router was not powered off but resetted 2 times: once to 6.49 and once to 6.48.3, and it start to work again as soon as I downgraded to the 6.48.3Ok that is interesting! This config works for me (it connects), but I have other strange problems with it. And downgrading does not fix it.
BTW, on some networks it may help to turn the router (or the PPPoE interface) off for a couple of minutes and then try again.
Because of that, you may get confused when actions like downgrading fix it, but just turning the router off for some time might have fixed it as well.
auto upgrade is disabled. As soon as I can do maintenance again I'll try to do as you suggest, thank youThe updating of firmware depends on the setting under system->routerboard->settings. I advise to keep that "auto upgrade" setting OFF and do the upgrade at your own convenience. That keeps you away from bricked devices due to unwanted firmware upgrades to broken versions.
Just rebooting the router may not take long enough to solve the issue with PPPoE setup I observe in our national carrier network. I advise to keep the router off for 5 minutes before concluding that the connection cannot be made.
so now I'm on 6.49 (until reboot) with 6.48.3 packages? is that possible?No this means you upgraded the firmware from version 6.49 and not (yet) from 6.48.3 (or you did not yet reboot after that).
You could (and should) have skipped that...RouterOS upgrade is fine. Its the routerboot firmware bit that screws it.
Tried again in the maintenance window but keeping it shutted down as you suggested didn't work.There is no need to update it every time nor to downgrade it now. The firmware is just the bootloader that initializes the hardware and loads RouterOS.
Originally it had a separate version number that changes only very occasionally, now it always has the same version number as the RouterOS it comes with, but usually there are no changes other than the version number. This time there appear to be changes (see the many troubles reported) but there is no changelog entry that describes a useful change in the firmware, so it would have been better not to upgrade it at all.
However, my RB4011 currently runs with RouterOS v6.49 and firmware v6.49 and I do not see the issues described above.
I also have a PPPoE connection over VLAN same as you describe (and with RFC4638 so 1500-byte MTU) and it works.
routerboard: yes
board-name: hAP ac^2
model: RBD52G-5HacD2HnD
serial-number: E5780DC.....
firmware-type: ipq4000L
factory-firmware: 6.45.9
current-firmware: 6.49
upgrade-firmware: 6.49
I came here looking for this. I recently pulled an RB750 out of storage to use temporarily. I updated the OS but not the firmware and it seemed fine until I did a /system reset-configuration. Now it's stuck in a boot loop, flashing all port LEDs every 10-20 seconds, and neither a 5-second config reset nor 10-second netinstall boot helps.hi,
Don't post often but a HEADS up for Mikrotik here.
We have had a collection of RB750 / RB951 based routers brick after upgrading to 6.49 Ros and Firmware.
Sadly we have been going around to customers to sort these router failures. Calling 6.49 Stable is a misnomer and shows a distinct lack of pre-release testing.I came here looking for this. I recently pulled an RB750 out of storage to use temporarily. I updated the OS but not the firmware and it seemed fine until I did a /system reset-configuration. Now it's stuck in a boot loop, flashing all port LEDs every 10-20 seconds, and neither a 5-second config reset nor 10-second netinstall boot helps.hi,
Don't post often but a HEADS up for Mikrotik here.
We have had a collection of RB750 / RB951 based routers brick after upgrading to 6.49 Ros and Firmware.
Hi,I updated the firmware on a hPA ac2 without issue. But your mileage may vary.Code: Select allrouterboard: yes board-name: hAP ac^2 model: RBD52G-5HacD2HnD serial-number: E5780DC..... firmware-type: ipq4000L factory-firmware: 6.45.9 current-firmware: 6.49 upgrade-firmware: 6.49
I can confirm !
This seems to be affecting Mipsbe hardware specifically.
For the RB951Ui, it look like the update is done correctly, but the package get damaged ( /system check-installation ) and if you reboot the device or there is a power outage ( no need to update the RouterBoard Firmware ) it's done ... It will just reboot in loop and you will have to use NetInstallI've update a RB951Ui and when update the Routerboard it reboots constantly in 30 seconds and i can't acces remotely. Is it there a bug on its Routers with this firmware ?
Had problem on my side with all MIPSBE board ( Mostly RB751 and RB951 ) ... upgrade to 6.49 work ...
BUT don't try to reboot the device or have a power failure else you just have a doorstop you have to flash with NetInstall.
If you run " /system check-installation " on MIPSBE and get " damaged routeros-mipsbe ... upload routeros-mipsbe-6.48.4.npk on your device and do a DOWNGRADE
After the downgrade I have been able to upgrade back to 6.49 and this time it work correctly and a reboot / power outage don't brick the device anymore !
Same here, got quite a few RB600 boards. I asked support about ROS 7 on RB600 and got following reply:Me too. I would really like to get ROS 7 on the RB600 fixed. It has plenty of storage and RAM, and the CPU is still very capable.
I also have RB800 and that runs ROS 7 just fine, both are same architecture, so it can't be that hard to make it work... I guess we will see what happens.The device might be too old to use with v7. Please test with any other or use CHR.
Oops.. sorry, I didn't see that. Sorry.But this is 6.49 thread, so enough of this off-topic ROS 7 mumbling...
The Long term version is affected too, as i mentioned in a post, there should be a "known stable" channel that contains the last version without major problems and th elog for it needs to show what small issues have been found with it over time since release.Well, in any case, and especially when equipment installed at customer locations is involved, you need to read the release topic of this forum before doing any upgrade.
No matter if it is called stable, .0 releases have usually experienced major problems in the past, and careful users have waited for the .1 or .2.
Finally I upgraded the last devices without issuesI upgraded a bunch of mipsbe devices to 6.49
- hAP AC (RouterBOARD 962UiGS-5HacT2HnT)
- CRS125 (CRS125-24G-1S)
- RB2011 (RB2011UAS)
- PWR-Line
upgrading firmware went fine ...
RB1100 (PowerPC)
hAP mini (RB931-2nD) (smips)
also went fine
waiting with my CCR1009 until there is more known about the reboot-delays ...
As mentioned in the long term thread >The issue with unavailable devices after 2nd reboot has been reproduced and it seems to be related when upgrading from 6.41.4 (or older) RouterOS/RouterBOOT versions. Our apologies for the caused issues, the fix will be included in upcoming releases.
So anyone using older than 6.42 RouterOS or RouterBOOT versions, please first upgrade RouterOS and RouterBOOT to 6.48.4 or 6.47.10, and only then use the latest version.
Is there anyone with long boot delay issue on CCR1036 (or any other Tile)? After flashing firmware 6.49 (RouterOS 6.49) we have very long boot time with our router.
Check installation is OK. Router boots but it takes about 8minutes or so.
We have managed to reproduce the issue locally in our labs and look forward to fixing it on upcoming RouterOS versions, unfortunately, I cannot provide an ETA now. Meanwhile please use 6.48.x.
upgraded RB2011 from 6.48.3 to 6.48.5 and after second reboot (firmware upgrade) it has gone in netboot mode.The issue with unavailable devices after 2nd reboot has been reproduced and it seems to be related when upgrading from 6.41.4 (or older) RouterOS/RouterBOOT versions. Our apologies for the caused issues, the fix will be included in upcoming releases.
So anyone using older than 6.42 RouterOS or RouterBOOT versions, please first upgrade RouterOS and RouterBOOT to 6.48.4 or 6.47.10, and only then use the latest version.
Was it running firmware 6.48.3 before, or an older firmware? Above, MikroTik says that this happens if the device is running 6.41.4 or older before the upgrade, if it is running newer firmware it is supposedly fine.upgraded RB2011 from 6.48.3 to 6.48.5 and after second reboot (firmware upgrade) it has gone in netboot mode.
Personally I disagree on this, I always do the RouterBOOT upgrades. MikroTik says the issue only appears here when the RouterBOOT upgrade is done from a very old version, 6.41.4 or older. We have upgraded quite a few devices successfully with no issues because we always did the RouterBOOT upgrades previously, and so all of the RouterBOOT firmware versions were new enough to not have this problem. But I am always really cautious with new RouterOS versions in general, and don't necessarily assume that they work properly or are safe to deploy. We always test them in our lab before doing an initial pilot deployment, then the final rollout.You could (and should) have skipped that...
Well, I wrote that in the context of the current issue, which was quite apparently caused by the RouterBOOT upgrade and not the RouterOS upgrade, and people not doing the RouterBOOT upgrade were not affected.Personally I disagree on this, I always do the RouterBOOT upgrades. MikroTik says the issue only appears here when the RouterBOOT upgrade is done from a very old version, 6.41.4 or older.You could (and should) have skipped that...
We don't upgrade RouterOS often. Our last standard version was 6.44.5 and early this year we adopted 6.47.9 as the new standard version after a lot of testing, so there are almost always guaranteed to be changes in the boot loader in that time.and extra wear. it is nonsense to align ROS and bootloader version whrn changes are so infrequent.
I don't completely agree with this - the problem with the old RouterBOOT versioning was it was different for different devices, and became hard to track what changes and versions were associated with what releases to determine whether an upgrade was necessary.Mikrotik going from proper RouterBOOT versioning to making it always the same as ROS version was the worst decision ever. There are rarely any RouterBOOT related changes in a changelog, so it's likely that only change is version number string. This stupid system also completely hides any importance of updates, because they all look the same.
Enough of this mess! Bring back the proper RouterBOOT versioning!
I had the same problem after upgrade "RB SXT 5HPnD"On SXT 5HPnD upgrading was OK but after "system routerboard upgrade" bootloader is broken (reboots constantly, no even single beep). Only netinstall helps!
It happened on two devices, didn't try more.
LOL dude Unprofessional is they should tested before it is relisted, anyway device work after upgrade but wont boot after second reboot.Unprofessional is to upgrade like 10 devices and then find out there is a problem... a month after release and without reading other user's experiences.
You have to know that "Stable" means that it is a version without ongoing experiments and regular changes, as opposed to "testing" and "development" where there are regular changes and additions.link https://mikrotik.com/download it says 6.49 (Stable) kinda not really.
I am aware I wrote about a different issue, but your post did remind me it exists. 😜@eworm I don't know if that was in reply to my message #232 but I am not discussing logging on RouterOS, merely the sending of DNS lookups for cloud.mikrotik.com in what appears to be a tight loop. When it fails, and a re-try does not fix that, it should not keep hammering.
My biggest gripe with netinstall is devices with multiple network interfaces, theres no way to select whch network interface to bind to so i have to manually disable all but my ethernet cross fingers and relaunch netinstall and hope to god it finds the primary ethernet which is a royal pain in the bum, i wasted an hour on this the first time and got caught out with it again later on.Netinstall is "black magic". When you rarely use it, it is always difficult to get it going.
Amongst things I encountered:
- no IP on host computer because it is setup for DHCP (has to have fixed IP)
- problem because ethernet goes down-up and computer takes too long to get networking running (solved by putting small dumb switch between computer and router)
- having a SUBST'd driveletter on the computer that points to some folder on C: completely confuses netinstall and makes it think you want to install a PC router
- non-intuitive user interface which often leads to wrong user actions
After a lot of fiddling I usually get it working but it is certainly not the fight you want to have with your only router that gives you internet access.
Always have a toy device ready to practice so you can be sure you are doing everything right before attempting to use it on something that won't boot!
Yes please. I also have >10 network interfaces on my laptop (bridges, vms, docker, VPN, etc) and disabling all of them except the one netinstall should use is PITA.Would it be so hard to add a drop box with all the network interfaces in it like TFTP32 and TFTP64 do?
That way I dont have to cripple my VPN adapters, bridge devices, VirtualBox, Wifi, WWAN adapters etc and then remember to enable them all again manually.
Yes, I have same experiences. So +100 votes for this request. But this is not thread for feature requests.My biggest gripe with netinstall is devices with multiple network interfaces, theres no way to select whch network interface to bind to so i have to manually disable all but my ethernet cross fingers and relaunch netinstall and hope to god it finds the primary ethernet which is a royal pain in the bum, i wasted an hour on this the first time and got caught out with it again later on.
Would it be so hard to add a drop box with all the network interfaces in it like TFTP32 and TFTP64 do?
That way I dont have to cripple my VPN adapters, bridge devices, VirtualBox, Wifi, WWAN adapters etc and then remember to enable them all again manually.
Hi,Hi guys,
i've bricked two CRS109-8G-1S-2HnD-IN. There is a solution to flash these device again actually? I've tried with netinstall etc without results.
And what about the updates so? I'm using manual upgrade to 6.47.10 :)