Hardware crash!

[schroeppel]

Hi,

is there a way to can view or extract emails in scalix 9.4.0 via comand line.
after a hardware crash is the system not realy runing.

Thanks for your help
Andy

[florian]

not easily. you will have to see what's going on.

are your scalix processes running? is the filesystem in sync and in order? did you run omscan to check your message store?

Florian.

[schroeppel]

the processes are running.
user get no acsess via webinterface. somtimes work the outlook client. and not all emails and contacts are displyed.

Andy

[florian]

have you run omscan to check the integrity of the message store?

omscan -Aafvx is what you want to get through. If that produces unrecoverable errors, your message store is corrupt.

Florian.

[schroeppel]

this show some errors like

omscan : [OM 4951]
A serious error has occurred. Please see the log files.
Event logged on 05.31.07 at 06:55:09.
Owner/Context Info : ~/user/g00003e
Additional error info:
omscan : [OM 3541] Content Record has not been upgraded to current container format.

omscan : [OM 4951]
A serious error has occurred. Please see the log files.
Event logged on 05.31.07 at 06:56:19.
Owner/Context Info : ~/user/g00003h
Additional error info:
omscan : [OM 28879] Bad magic number in an item index record.

thanks
Andy

[florian]

jup. these are signs of significant message store corruption caused by the crashed filesystem. Walking through further diagnosis of thisis beyond what i think we can do in a forum; you might or might not be able to rescue data from here. i think you will need to purchase/open an incident with scalix tech support who can try to see what can be rescued. obviously, the other way is to restore to the last backup.

Florian.

[schroeppel]

Florian,
Thanks for the relpy. what can i restore without more problems?

Andy

[florian]

well, in the case of a crash, you would typically want to clear out and recreate the filesystem that holds /var/opt/scalix and then restore the whole tree from backup. There are numerous interconnections between these files.

If you have done user-level backup before the crash, i.e. using sxmboxexp/omcpoutu, you could also try deleting users until the omscan errors go away and then re-adding them and importing the data from the user-level backup files. Under normal circumstances, even the worst case of message store corruption will normally be limited to a single user account or a couple of accounts, but rarely affect the whole system.

the error messages referenced point to two users, the number in the g-directory name tells you which, e.g. 3e (this is base 32) = 3*32 + e = 96 + 14 = 110 and 3h = 3*32 + h = 96 + 17 = 113. omshowu -O <number> should show you the user record, if I am not mistaken. (typing this without testing or looking at the manpages)

Florian