Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 24 Mar 2006 12:58:09 +0100 (CET)
From:      "Julien Gabel" <jpeg@thilelli.net>
To:        "Kris Kennaway" <kris@obsecurity.org>
Cc:        Eugene <genie@geniechka.ru>, freebsd-fs@freebsd.org
Subject:   Re: UFS2 Snapshots in 6.1-Beta4 - Confirmed Problems.
Message-ID:  <59591.145.248.192.4.1143201489.squirrel@webmail.thilelli.net>
In-Reply-To: <20060323213941.GA88109@xor.obsecurity.org>
References:  <20060322120033.E5B8916A44B@hub.freebsd.org> <007801c64eb9$3db4c050$1b12c055@genie> <50319.192.168.1.101.1143148607.squirrel@webmail.thilelli.net> <20060323213941.GA88109@xor.obsecurity.org>

next in thread | previous in thread | raw e-mail | index | archive | help
>>> While we're on the topic of snapshots, I recently posted this
>>> follow-up to the PR kern/66270 "[hang] dump causes machine freeze"

>>>> I wonder what is the status of this PR?
>>>> I seem to have the same problem with 5.3-RELEASE.
>>>>
>>>> Before this started, system worked nicely for more than a year and
>>>> then it started to hang during dump (seemingly just after mksnap_ffs)
>>>> first intermittently and now reproducibly (might increase in data
>>>> size be a factor? though disk is only about 18% full).
>>>> Originally it hanged and console had complains about increasing
>>>> PMAP_SHPGPERPROC and maxproc.
>>>> After some talk in freebsd-fs, I adjusted these parameters and now
>>>> it just hangs without complaining =((
>>>>
>>>> Any suggestions?

>>> In addition to this PR, I found a couple of similar threads in
>>> different lists but they all just have some initial talk and
>>> preliminary analysis, and then it just stops. Was that snapshot
>>> problem fixed or is it the same as we are talking now? I can't
>>> believe I am the only one in almost 2 years who is (trying to be)
>>> doing automated backups...
>>> So, if it was fixed, when (which version)? If not, then when it might
>>> be and -- in the meantime -- is there any workaround for backups?
>>> (The system is the production server on the colocation so doing
>>> backups in single-user mode is not practical).

>> Interesting.  I used dump/restore without problem on i386 using a file
>> system of 120Go.  I swiched two weeks ago on amd64 using a file system
>> of 260Go and encountered the very same behavior (hang, freeze).  Don't
>> know if it is related to the new processor architecture or the bigger
>> fs size, but i didn't observe this hang on smallest file system before.
>>
>> Because i am a big fan of dump, i currently choose to use it without the
>> '-L' switch, back to the old (classical?) behavior, i.e. not doing a
>> snapshot before doing the dump.  It is always better than not having
>> complete system backup at all (YMMV, though).
>>
>> I am currently running 6.1-PRERELEASE build Wed Mar 22 03:55:42 CET
>> 2006.

> Let us know if it recurs with your updated kernel.

Well, it seems ok now with FreeBSD 6.1-PRERELEASE #0 Fri Mar 24 03:58 CET
2006.  During the creation of the snapshot (either with mksnap_ffs(8) or
with `dump -L'), the system was not particularly "responsiveness", but it
seems to be the expected behavior.  FYI, for an UFSv2 file system of
260Go with currently ~6.5Go used, it took approximatively ~15mn (on
recent hardware).

Globally, and minus the ~15mn of not very good serviceability (http
requests time-out -- or very, very slow to be served -- and interactive
shells not very responsive), all seems ok : the snapshot doesn't hang or
freeze the system globally anymore, and the backup are made correctly
using `dump -L').  So, good news.

-- 
-jpeg.




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?59591.145.248.192.4.1143201489.squirrel>