Hi all,
I am running
Solaris Express Community Edition snv_130 X86
Copyright 2009 Sun Microsystems, Inc. All Rights Reserved.
Use is subject to license terms.
Assembled 14 December 2009
with 2 pools, rpool (installed as version 22) and brick (upgraded to version
22).
Yesterday I had a scenario where I had a hard hang. I powered cycled the system
and all looked OK.
Later I noticed that some logfiles had been truncated ( and the modify
timestamp had been set to 1 minute into the reboot)
and that some files that had been moved and renamed, where back the there
previous state (about 4 GB of files). These changes had been made
40-60 minutes before the hang. So I was expecting that the writes would have
been committed to disk.
It looks like that some files systems have rolled to a previous state (I though
this would happen at a pool level) .
I can find no logging that and old uberblock was used.
It looks like that during the import we had to rollback to an old uberblock.
So my questions are:
1) Does this scenario make sense?
2) How long should it be before the writes are committed to disk?
3) Should this sort of recovery be be @ a fs or pool basis?
4) Is this type of rollback logged any were?
Regards
Rodney
_______________________________________________
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
