Networker

Re: [Networker] Slow mminfo response

2009-06-19 16:06:12
Subject: Re: [Networker] Slow mminfo response
From: Tim Mooney <Tim.Mooney AT NDSU DOT EDU>
To: NETWORKER AT LISTSERV.TEMPLE DOT EDU
Date: Fri, 19 Jun 2009 14:59:44 -0500
In regard to: [Networker] Slow mminfo response, Stan Horwitz said (at...:

I have two NetWorker servers. Let's call one X and the other server Y.
Server X is a Sun T2000 with Solaris 10 and NetWorker 7.4.1. Server Y is a
Dell 2950 with Red Hat Linux AS 4.5 and NetWorker 7.4.4. The Hobbit server
is also running Red Hat Linux and NetWorker client 7.4.4.

On the Hobbit server, when I do something like

mminfo -v -s X -c client1 -ot

The response takes around 6-7 minutes to appear.

If I do something like

mminfo -v -s Y -c client2 -ot

I get a near instantenous response.

If I log onto server X and do the same

mminfo -v -s X -c client1 -ot

Command, I get a near instantaneous response.

NetWorker server X has two NICs. One NIC has various tagged VLAN addresses
associated with it. They all work fine for backups. The second NIC is the
one over which the mminfo query is being done from the Hobbit server. This
second NIC isn't in heavy use at all. Its 1GB and its autonegotiating
properly with our switch.

If I were in the position of having to debug this, I would reach for my
favorite debugging tool of all time: truss (strace on Linux).  You can use
truss on Solaris and strace on Linux to debug what's happening on both
ends of the connection, while it's taking place.  You need to be
comfortable reading system call traces, but if you are, you're pretty much
guaranteed to be able to find the problem.

I have been working with one of our network management type people on this
problem and he sees nothing wrong on either end. Server X responds to mminfo
commands quickly on a variety of other servers, except when the commands
come from our Hobbit server. The mminfo commands to NetWorker server Y also
result in fast responses. I can also back up the Hobbit server to Server X
reasonably fast.

Ok, start by debugging just the mminfo command run from your hobbit
server, using strace.  First try one against server Y, the good one, and
then try one against server X.  You may want to use one of the timing
arguments (e.g. -r) to strace, so you also have timing information about
how long each call is taking.

If that doesn't give you enough info, you can also simultaneously debug
what the daemons are doing on each of the NetWorker servers.  Just use
truss/strace to attach to the running pids of the relevant daemons and
record the output to a file.

Tim
--
Tim Mooney                                             Tim.Mooney AT ndsu DOT 
edu
Enterprise Computing & Infrastructure                  701-231-1076 (Voice)
Room 242-J6, IACC Building                             701-231-8541 (Fax)
North Dakota State University, Fargo, ND 58105-5164

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 with 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

<Prev in Thread] Current Thread [Next in Thread>