SwOS 2.18 reboot every 23 minutes

SwOS
1

Hi!

We have several CRS312 installed.

At the office, one CRS312/r2 (v2.17) and one recent CRS312/r3 (v2.18).

We also are currently installing 2 new CRS312/r3 on a customer site, both v2.18.

All switches have their 2 power supplies plugged in.

What I see at customer site is that each and every 23 minutes, both CRS312 reboot. I switched one of them off then on again, to desynchronize them, they reboot after their “own” 23 minutes, reboots are not caused by an event common to both switches. When I say 23 minutes, it’s 23 minutes “sharp”, the uptime counter on the system tab stops increasing at 23:01. Then the switch reboots.

I tried many things, down to kind of a minimum architecture with a single switch, same issue.

At that point, I connected to our office, and checked how it was going there. And I saw that :

  • the CRS312/r2 v2.17 had more than 91 days of uptime

  • the CRS312/r3 v2.18 had only a few minutes

So I observed the /r3 one, and saw that it rebooted too! But every 24 minutes, not 23! And 24 minutes sharp, uptime counter at 24:01.

I made some searches on the net, nothing such…

Any idea of what I could do to fix that ? Having no logs doesn’t help much…

Also, I don’t see the watchdog checkbox within the System tab on the three /r3 (v2.18), but I see it on the /r2 (v2.17). Is it the 2.18 that does that ? I don’t see anything such in the release notes…

Thanks by advance!

Regards,

Pascal.

2

I had the same problem. It appeared after I disabled the ETH/BOOT port. Exactly every 23 minutes. I re-enabled the ETH/BOOT port and the problem disappeared.
Wonder if this might be the case for you as well?

3

Has anyone ever created a support ticket about this ?
What if you use ROS, not SWOS ? Does the same happen ? (to possibly rule out if it is OS related or HW related)

4

Yes, I opened a support ticket last Thursday (SUP-203081).
I’m on an CRS312-4C+8XG r2 SwOS v2.18.
Haven’t tested with ROS and since it’s a production switch I’m not very eager to do so.
I hope Mikrotik can reproduce the issue.

5

Hi Marc!

What you said was sounding like black magic, I don’t really understand how both things might be related, but the thing is, it fixed the problem! :slight_smile: Including my local switch that used to reboot after 24:00 minutes, not 23:00…

You avoided me a 1500kms trip, man, I owe you one beer! :wink:

@holvoetn : I had opened a ticket quite some days ago, Mikrotik support got back to me 10 days ago asking me to capture all the config screens, which I actually did, but I now realize that I forgot to send them back… My bad! I’ll do so then, telling it’s fixed.

And in the meantime, I had switched to RouterOS trying to obtain an equivalent config, and I can tell you I see no such behavior when the switch is running RouterOS, the problem is definitely at the SwOS level!

Both have a nice evening, and thanks for your help!

Best regards,

Pascal.

6

Hi Pascal,

Glad that I could be of help. I updated my support ticket with a link to this thread, maybe you can do the same.

Hope it gets fixed soon. I guess not a lot of people run with the ETH/BOOT port disabled otherwise the sh*t would've hit the fan earlier.

Have a nice evening too!

Marc

— edit, can’t post more than 3 messages in the same topic as a new user
@holvoetn maybe you can mod this to be a new reply?

–>

@pascal.FR I just got a reply back from support:

We were able to reproduce your issue in our labs.
Our developers have already considered this issue, and it will be fixed in future releases, but unfortunately, I cannot share any ETA yet.

7

To be fair, It still looks like magic, only It Is definitely white magic.

8

You should be able to do so ?
Anyhow, quoted your entry so it appears as a new post. But unfortunately I can not make it yours.

9

Thanks!

10

Thanks, everyone, for reporting this!
We manged to reproduce it and looking forward to fixing it in future SwOS versions.