NW5.1 SP6 - high utilization
al...@canisius.edu
our Netware 5.1 SP6 servers. We had been running Symatec Antivirus Corp.
Edition 8.0 on these servers for some time with no problem. Since this was
the only software that was definite to change (and having seen some
isolated times of bad virus definition updates), I unloaded this from each
of the servers. Utilization would not decrease, but after a restart, 3 of
the servers seem to be OK for now. The 3 that are running well are all
Compaq Proliant ML370's with 1GB RAM each. I have one that is at a remote
site (connected via T1) that had been running trouble free for 140+ days
since last restart. This is a Compaq Proliant 1600 with 640M RAM, but is
also a light-duty assignment.
As with the other servers, I unloaded & commented out SAV CE from
starting, also commented out portal.nlm since I haven't been using it &
had some problems anyway. This server took 30+ minutes to start up and
continues to have high utilization (at or near 100%). I find no evidence
remotely of hardware failure. Looking at the top threads only shows server
processes & a few other things that appear somewhat normal. We use Legato
Networker, but no backups have been running while I still see high
utilization.
Looking at TID10018089, I'm a bit wary of trying the workthrd.nlm since
this is older.
We are also running eDirectory 8.7.0.4 currently & had flattened the tree
a bit during the winter break (we had student accounts broken into 8
subcontainers, now all are moved up to the same level - approx. 5700
objects).
What tools are available to see what is going on & where can I find these
in short order? From the long boot time early this morning, it almost
seems possible there is a hardware issue, but nothing has changed & no
errors are being logged anywhere to suggest this.
Thanks for any possible assistance/suggestions!
--alan
Joseph Moore [SysOp]
> This server took 30+ minutes to start up and
> continues to have high utilization (at or near 100%). I
>
hmmmm - have you check your free space on the SYS volume? If your free
blocks fall below 10%, it will go into "agressive suballocation" and
start suballocating to the exclusion of all else. If, as you say,
*nothing* changed on the box other than the AV updates, and it still is
showing high util, try running a purge on the sys and other volumes
Joe Moore
Novell Support Connection Volunteer Sysop
http://just.fdisk-it.com - Coming soon:
"Migrating From NetWare Enterprise Web Server to Apache 1.3.x"
http://www.caledonia.net/sysops.html
NO EMAIL PLEASE!!!!!
al...@canisius.edu
>
> > This server took 30+ minutes to start up and
> > continues to have high utilization (at or near 100%). I
> >
>
> hmmmm - have you check your free space on the SYS volume? If your free
> blocks fall below 10%, it will go into "agressive suballocation" and
> start suballocating to the exclusion of all else. If, as you say,
> *nothing* changed on the box other than the AV updates, and it still is
> showing high util, try running a purge on the sys and other volumes
>
The volume is only 34% allocated, but I purged this & the other volumes.
All volumes are on 1 RAID5 array using traditional NW filesystems.
Utilization has not changed; staying in the low 90% range with occasional
drops into the 70% range. The long term cach hits are at 5%, with an LRU
sitting time of over a day (about 1 minute short of the server uptime). I
just went through comparing set params from a copy of config.txt from last
June & other than the additional params either from SP6 or eDir 8.7.0.4
everything is the same. I still have not reloaded/reinstalled Symantec
Antivirus (not a pleasant thing with the recent worm activity).
--alan
Joseph Moore [SysOp]
> Utilization has not changed; staying in the low 90% range with occasional
> drops into the 70% range.
>
hmmmm - mabye LOAD DOSFAT.NSS (I think that is the command) will help?
al...@canisius.edu
>
> > Utilization has not changed; staying in the low 90% range with
occasional
> > drops into the 70% range.
> >
>
> hmmmm - mabye LOAD DOSFAT.NSS (I think that is the command) will help?
>
I appreciate your assistance with this, however I'm not sure how loading
an NSS module when I'm not using NSS for any volumes might help.
Joseph Moore [SysOp]
> however I'm not sure how loading
> an NSS module when I'm not using NSS
>
whoops - sorry about that, there is a utilization problem that
DOSFAT.NSS helps with - but as you say, you are not using NSS (though
it might not hurt to try it). Do you have any of the post SP6 patches
for TCPIP and NLS loaded? Have you run the PURGE_NW utility? If not,
do those and see if they help - if they don't, then please post
config.txt with public IP's or rconsole passwords obscured.
Andrew C Taubman
the registry out to the Dos partititon every 5 minutes, which flips the
server into real mode. On some hardware this causes high util. Loading
dosfat.nss allows the server to write to dos without flipping into real
mode. Definitely worth trying.
--
Andrew C Taubman
Novell Support Forums Volunteer SysOp
http://support.novell.com/forums
(Sorry, support is not provided via e-mail)
Opinions expressed above are not
necessarily those of Novell Inc.
Joseph Moore [SysOp]
> Loading
> dosfat.nss allows the server to write to dos without flipping into real
> mode. Definitely worth trying.
>
cool, thanks!
al...@canisius.edu
>
> > however I'm not sure how loading
> > an NSS module when I'm not using NSS
> >
>
> whoops - sorry about that, there is a utilization problem that
> DOSFAT.NSS helps with - but as you say, you are not using NSS (though
> it might not hurt to try it). Do you have any of the post SP6 patches
> for TCPIP and NLS loaded? Have you run the PURGE_NW utility? If not,
> do those and see if they help - if they don't, then please post
> config.txt with public IP's or rconsole passwords obscured.
>
tomorrow morning. I'm not familiar with the purge_nw utility, but will
snag that & give it a go. This morning I purged deleted files on all
volumes on that server; one took 1.75 hours to complete! Had ZfD3.2 sybase
database running, but no one was using it (inventory handled by another
non-Novell product), so I unloaded it. Have zfdstart.ncf commented out of
all 4 servers now.
And as Andrew suggested, I'll also load DOSFAT.NSS.
Thanks!!
--alan