I just wanted to give everybody a heads up on some issues we had after
upgrading to Networker 7.3. The issues are serious enough that we are
moving back to 7.1.4.
Our environment: Windows 2000 (reasonably up to date patches). Two
autoloaders, Overland/SDLT220 and Qualstar/LTO3. One diskbackup storage
array, 1.5TB. All devices are SCSI attached.
Our normal operations send backups to disk, which are then cloned to tape.
We have a script that removes savesets from the disks after they have been
cloned and when free disk space below a set threshold. The goal with this
is to keep as much data on disk so that the disks can be used for
recoveries in many cases without having to recall tapes from offsite
storage. This has been working ok under 7.1.3.
The upgrade to 7.3 was pretty painless, it installed pretty quickly and
without any problems. It ran for a few days without any problems.
After four days, when checking the status of the backups from the previous
night, we found that the managment console had been disconnected from the
server. We normally leave the managment console open on the server for
quick status checks and such. After closing the console, we were unable to
re-open it. We ended up restarting the server to clear the problem. After
restarting everything worked fine for a while. After a few hours the
system started to slow down and we found that the gstd.exe (probably the
server-side manangement console daemon) process had grown to 1GB of RAM.
We restarted that service only, which reset it back to a reasonable amount
of memory, but grew at a rate of 300MB per hour.
The next problem that we found was that after restarting the networker
server, one of the autoloaders was not working properly. Inventory
commands would complete successfully after 4 seconds without an actual
inventory being done (no barcode reader in the loader), or the inventory
being updated. We tried both gui and command line operations (nsrjb) and
they both acted identically. After talking with tech support, we decided
to remove the afflicted loader and re-add it, at which time it started
working properly. This happened four times, three times with the overland
loader and once with the qualstar. It seemed to be a random choice as to
which one would not work.
The next problem was the one that ultimately caused us to back down to an
older version. It appeared that the media database hung on two different
occasions. When this happened, mminfo, nsrim, and nsrmm commands would
hang (for hours), and the gui would hang during loading. I'm not sure if
the problem was with the media database, but that is where I saw the
effects of whatever the problem was.
The final problem that I ran into was that when doing manual clones (using
nsrclone) of savesets on the adv_file device, while other backups were
running, was that the first manual clone would run and complete fine.
After that first one, any other clones will wait because the "server is
busy". However, automatic staging policies were still working.
All of these issues were reported to tech support, although they were
appeneded to the low priority gui failure problem. I also believe that
some of the issues may be related to, or triggered by, our environment and
how we are using the software.
Hopefully 7.3.1 will work out better for us.
Thanks,
Erik
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
|
