[OmniOS-discuss] stuck in maintenace

Jim Klimov jimklimov at cos.ru
Sun Dec 18 19:01:26 UTC 2016


17 декабря 2016 г. 13:58:02 CET, Martijn Fennis <martijn at fennis.tk> пишет:
>Hi,
>
>
>During zfs rebuild, memory got used untill i had to unplug the host.
>
>Added the arc memory limit.
>
>After a second or third reboot, the box only returns in maintenance
>mode.
>
>
>Did some googling and found to :
>
>
>svcadm clear system/boot-archive
>
>but it returns with the same:
>
>svc.startd[10]: svc:/system/boot-archive:default: Method or service
>exit timed out.  Killing contract # number
>
>
>Is there a way to fix this?
>
>
>Thanks,
>
>Martijn
>
>
>------------------------------------------------------------------------
>
>_______________________________________________
>OmniOS-discuss mailing list
>OmniOS-discuss at lists.omniti.com
>http://lists.omniti.com/mailman/listinfo/omnios-discuss

This sounds reminiscent of issues I had some 6-7 years ago on opensolaris lists (back then). In particular, the seeming culprit was that zfs under-the-hood activities like large pool deletions needed more kernel ram than i had physically available, so after processing some queued operations, the box halted with no mem available (under 32M for several minutes in a row).

It took 2 weeks of reboots for the box to complete its queue of operations (which it did progress, monitorable with zdb -DD IIRC) and a nujber of worried emails to the lists ;)

Not sure your case is similar, but try to track any delayed operations (maybe nowadays via 'zpool get all rpool') to see if progress is being made across reboots. Maybe younjust heec to be patient...

Jim
--
Typos courtesy of K-9 Mail on my Samsung Android


More information about the OmniOS-discuss mailing list