Hi Scott,
I am not quite sure if I understand softlimit/hardlimit concept
in storagecheckpoints. According to manual:
> If a hard limit is exceeded, all further allocations on any of the
> Storage Checkpoints fail, but existing Storage Checkpoints are
> preserved. Without the -f option, one or many removable Storage
> Checkpoints are deleted to make space for the operation to succeed.
I have setup hardlimit:
kos:/var/tmp# fsckptadm getquotalimit /vxfs/test-vol
Filesystem hardlimit softlimit usage
action_flag
/vxfs/test-vol 31000 0 61611
I have the following storage checkpoints:
kos:/root# fsckptadm -v list /vxfs/test-vol
/vxfs/test-vol
f3:
ctime = Wed Sep 24 10:45:21 2008
mtime = Wed Sep 24 10:45:21 2008
flags = largefiles, removable
# of inodes = 32
# of blocks = 61441
# of reads = 0
# of writes = 0
# of pushes = 0
# of pulls = 0
# of moves = 0
# of move alls = 0
# of merge alls = 0
# of logged pushes = 0
# of enospc retries = 0
# of overlay bmaps = 0
f2:
ctime = Wed Sep 24 10:45:13 2008
mtime = Wed Sep 24 10:45:13 2008
flags = largefiles, removable
# of inodes = 32
# of blocks = 1
# of reads = 0
# of writes = 0
# of pushes = 0
# of pulls = 0
# of moves = 0
# of move alls = 0
# of merge alls = 0
# of logged pushes = 0
# of enospc retries = 0
# of overlay bmaps = 0
f1:
ctime = Wed Sep 24 10:44:48 2008
mtime = Wed Sep 24 10:44:48 2008
flags = largefiles, removable
# of inodes = 32
# of blocks = 1
# of reads = 0
# of writes = 0
# of pushes = 0
# of pulls = 0
# of moves = 0
# of move alls = 0
# of merge alls = 0
# of logged pushes = 0
# of enospc retries = 0
# of overlay bmaps = 0
It seems that one (a few ?) of them should be deleted to make
space. Could you please tell me more where is the problem ?
Regards
przemol
On Tue, Sep 23, 2008 at 03:57:38PM -0700, Scott Kaiser wrote:
> Hi Przemyslaw,
>
> Have you looked at the existing quotas?
>
> http://sfdoccentral.symantec.com/sf/5.0/solaris/manpages/vxfs/fsckptadm_
> 1m.html#429951
>
> The -r option doesn't delete any checkpoints, it just makes each
> individual checkpoint 'removable'. Plus, it's been the default since
> 4.0, so unless you have a really old system, you don't need to specify
> it.
>
> It's the setquotalimit that matters. That is expressed in absolute limit
> (# of fs blocks). Are you saying basically, yes, that's what you want,
> just allow it to be relative (90%)?
>
> Also note that the hard limit on checkpoints doesn't necessarily remove
> only the oldest checkpoint when exceeded -- it removes enough removable
> (again, default) checkpoints starting with the oldest until all
> checkpoints are below the hardlimit.
>
> Scott
>
> -----Original Message-----
> From: [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED] On Behalf Of
> [EMAIL PROTECTED]
> Sent: Monday, September 22, 2008 7:02 AM
> To: veritas-vx@mailman.eng.auburn.edu
> Subject: [Veritas-vx] storage checkpoints - quota
>
> Hello,
>
> we are evaluating storage checkpoints. Currently we are testing how does
> it cope with full filesystem.
> The '-r' option is really usefull (since it removes the oldest storage
> checkpoint when there is no space) but I miss very much an option which
> will remove such sckpt when e.g. filesystem reached 90%.
> Are Symantec's engineers evaluating such option ?
> It is usefull when you have a lot of filesystems and setting
> (calculating) quota for each filesystem is time consuming (especially in
> a dynamic environment).
>
> Regards
> Przemyslaw Bak (przemol)
> --
> http://przemol.blogspot.com/
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> ----------------------------------------------------------------------
> Kredyt Hipoteczny w Banku Millennium - zdobywca Zlotego Lauru Klienta!
> Sprawdz >> http://link.interia.pl/f1f15
>
> _______________________________________________
> Veritas-vx maillist - Veritas-vx@mailman.eng.auburn.edu
> http://mailman.eng.auburn.edu/mailman/listinfo/veritas-vx
--
Regards
Przemyslaw Bak (przemol)
--
http://przemol.blogspot.com/
----------------------------------------------------------------------
Doladuj w promocji telefon!
Sprawdz >> http://link.interia.pl/f1f18
_______________________________________________
Veritas-vx maillist - Veritas-vx@mailman.eng.auburn.edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-vx
