Veritas-bu

[Veritas-bu] hanging backups.

2006-08-17 12:23:43
Subject: [Veritas-bu] hanging backups.
From: jlightner at water.com (Jeff Lightner)
Date: Thu, 17 Aug 2006 12:23:43 -0400
Is the master a Windows machine as well?  If its Unix/Linux you can
download and install tcpdump which will let you specify interface and IP
you wish to monitor.  You can use ethereal to look at the packets
captured by tcpdump.

Not sure if there's a Windoze equivalent or version of those tools.

-----Original Message-----
From: veritas-bu-bounces at mailman.eng.auburn.edu
[mailto:veritas-bu-bounces at mailman.eng.auburn.edu] On Behalf Of Paul
Keating
Sent: Thursday, August 17, 2006 11:19 AM
To: veritas-bu at mailman.eng.auburn.edu
Subject: [Veritas-bu] hanging backups.

So I've got about 6 machines, all MS-Windows at a remote site connected
by 1000Mb/s ethernet over DWDM/Fiber.

It seems on a nightly basis, it rotates, which of thes machines hang.

The behaviour looks like the stereotypical NIC duplux mismatch, but
that's been verified, and is not the case....no CRC errors or anything
along the way, and the hang doesn't happen every night.....if a machine
hangs one night, it might be fine for a week, but others will take turns
having a bad night....on the bright side, I don't believe we've missed
the same machine twice in a row.

the server on this side and the client on the other side are connected
directly to switches, which are then connected directly to the DWDM
nodes....duplex and speed has been verified all along the way...

We've been sniffing traffic trying to capture something, but getting the
right client on the right NBU server interface is like playing
whack-a-mole....a couple nights ago, we caught one, and got the
following trace:

Source    Dest    Summary
master    client    DSI: Continuation of missing frame; 4 bytes of data
master    client    DSI: Continuation of missing frame; 1 byte of data
master    client    TCP: D=814 S=865    ACK=4282441235    WIN=64240
master    client    DSI: Continuation of frame 41; 4 bytes of data
master    client    DSI: Continuation of frame 41; 1 bytes of data
master    client    TCP: D=814 S=865    ACK=4282441236    WIN=64240
master    client    DSI: Continuation of frame 41; 4 bytes of data
master    client    DSI: Continuation of frame 41; 1 bytes of data
master    client    TCP: D=814 S=865    ACK=4282441237    WIN=64240

The "missing frame" message at the beginning and the repeated
"continuation of frame 41" got me intrigued, but I don't really know
what to make of it.

???

The interesting thing is that the linux and solaris boxes at the remote
site have been 100% problem free.

Paul


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