I have a number of adv_file devices in the same pool, they're all set up
with a target parrallelism of 1 which effectively forces a round-robin
effect. Each filesystem is 2.5TB ufs and I have a rule that no saveset
should be larger that 250GB (that's the hard bit). I have arranged
things so that when utililisation of a particular filesystem gets above
80% it will have data deleted until utilisation gets down to 50%, so we
don't spend much time above 80%, and we delete in big chunks at a time..
I'm not very happy with this arrangement as I don't see why I have to
waste 20% of my space to avoid full disks and/or filesystem
fragmentation, but it works most of the time.
Where this usually goes wrong is that by chance a set of large savesets
all go to the same volume and it can fill. I can work around this
incident by putting that volume into service mode, and killing the
nsrexecd processes for as many savesets as I need to - the ones that get
killed have to restart (group retrys value needs to be more than 1), and
when they restart they go to another filesystem. (this may not be
possible in 7.3)
I have experimented with using just one huge filesystem (8TB) for a pool
and using subdirectories as the adv_file volumes - this way you have all
your space available no matter which volume is chosen, and you can
create as many subdirectories as you have tape drives, it works, however
you are then causing a lot of random read/write activity all over the
filesystem and ufs and vxfs both get fragmented very quickly, and the io
pattern puts a lot of load on raid-5 or raid-3 backends.
I have had some success in the lab with sun QFS, I am also looking at
zfs currently - you can fight the fragmentation with these filesystems
by choosing a large allocation unit - eg 10 or 100 meg, or maybe even a
gig.
I know less about linux - but you have several different f/s choices
like reiser-fs which may help with the issues around huge filesystems
too.
-----Original Message-----
From: Legato NetWorker discussion [mailto:NETWORKER AT listserv.temple DOT edu]
On Behalf Of Rich Graves
Sent: Tuesday, 11 April 2006 5:16 AM
To: NETWORKER AT listserv.temple DOT edu
Subject: [Networker] Backups to adv_file can't span volumes?
I have 4 1.5 TB adv_file devices. It's split up that way because Linux
2.4 isn't terribly happy about > 2TB physical volumes, and because it
affords some parallelism in cloning from adv_file to our 4 tape drives.
One volume filled up, and I was surprised to find a backup hanging on
"waiting to free adv_file space..."
Is there any way to convince NetWorker (7.2.1, Linux) to continue the
save onto an adv_file device in the same pool that has room, as it would
do with tapes?
To sign off this list, send email to listserv AT listserv.temple DOT edu and
type "signoff networker" in the body of the email. Please write to
networker-request AT listserv.temple DOT edu if you have any problems wit this
list. You can access the archives at
http://listserv.temple.edu/archives/networker.html or via RSS at
http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER
NOTICE
This e-mail and any attachments are confidential and may contain copyright
material of Macquarie Bank or third parties. If you are not the intended
recipient of this email you should not read, print, re-transmit, store or act
in reliance on this e-mail or any attachments, and should destroy all copies of
them. Macquarie Bank does not guarantee the integrity of any emails or any
attached files. The views or opinions expressed are the author's own and may
not reflect the views or opinions of Macquarie Bank.
To sign off this list, send email to listserv AT listserv.temple DOT edu and
type "signoff networker" in the
body of the email. Please write to networker-request AT listserv.temple DOT edu
if you have any problems
wit this list. You can access the archives at
http://listserv.temple.edu/archives/networker.html or
via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER
|
