CCR one beep of doom after power cut

Been rolling out CCR’s since 2013 and other than the odd crash, been very impressed with reliability, have about 12 out in the network now, all 1009 and 1009 +.

Long story short, one of our sites with the oldest 1009 CCR had a power cut. It has a APC UPS that ran for about 2 hours then shutdown (the CCR is not connected to the APC for signalling so the shutdown was ‘dirty’ for the CCR).
When power came back CCR didn’t boot successfully. It powered fine, beeped once but then nothing. When I got a console cable on it, I could see it was having issues unlocking files, it was like the flash had gone read only.

I managed to recover it using BootP and it seems fine now. But I wonder if this a freak thing, or should I expect the others to start to fail like this. Be interested in others experiences of power cut failures of CCR and would you re-deploy a unit after it did this?


Thanks

W

Did you netinstall it?

Does it have logging to disk, user-manager or or graphing?

If its NAND was being accessed I’d say it’s normal for the filesystem to become corrupted.

I’d netinstall i and check for bad blocks percentage.

Does it reboot fine?

Yep I did netinstall it. Once I got my head around the BootP IP had to be one in the same subnet, and not the PC running the netinstall program, it worked fine!

It didn’t have graphing, or any type of logging to disc.

I will read up on bad blocks percentage and check it.

It does reboot and appears completely normal, since I netinstalled it.

Thanks

W

Sometimes NAND filesystem gets corrupted on blackouts (usually when they’ve happened repeteadly), causing these kind of problems, have happened to me in the past and so far most devices are performing fine after a netinstall.

OK thanks. I will test this unit thoughly before I redeploy it.

Is there any way to check for bad blocks? System > resource doesn’t display it anymore.

http://forum.mikrotik.com/t/why-mikrotik-needs-shut-down-properly/89177/1