Networker

Re: [Networker] Exchange Restores Hang

2004-01-01 20:11:13
Subject: Re: [Networker] Exchange Restores Hang
From: "Eichelberger, Jon" <jon.eichelberger AT SAP DOT COM>
To: NETWORKER AT LISTMAIL.TEMPLE DOT EDU
Date: Fri, 2 Jan 2004 02:10:59 +0100
Hi, Scott.

I guess your meaning of counter-intuitive is "only useful to someone
with a copy of the C code".  I used a "-D 5" on nsrxchrc in
a DOS command session.  When the hang happened, here's what
was at the end of the file to which I was logging the debug output:

[a whole lot of the same 3 lines (see below) ending with...]
nsrxchrc: rcutil.c(511): Calling _nwbsa_getdata32()
nsrxchrc: rcutil.c(519): Received 32768 bytes of data
nsrxchrc: rcutil.c(533): Calling rHandle->write_file()
nsrxchrc: rcutil.c(511): Calling _nwbsa_getdata32()
nsrxchrc: rcutil.c(519): Received 32768 bytes of data
nsrxchrc: rcutil.c(533): Calling rHandle->write_file()
nsrxchrc: rcutil.c(511): Calling _nwbsa_getdata32()
nsrxchrc: rcutil.c(519): Received 32768 bytes of data
nsrxchrc: rcutil.c(533): Calling rHandle->write_file()
nsrxchrc: rcutil.

Note the incomplete last line.  I have not aborted the restore yet and I started
it over 24 hours ago.  The restore is still sitting there hung.  I'm hoping 
whatever might be in
the output buffer will get flushed when I abort the restore tomorrow.

Regards,
   Jon

-----Original Message-----
From: Scott Bingham [mailto:SBingham AT legato DOT com]
Sent: Wednesday, December 31, 2003 3:28 PM
To: 'Legato NetWorker discussion'; Eichelberger, Jon
Subject: RE: [Networker] Exchange Restores Hang


Hello Jon,

There were a couple of reasons why the NetWorker Server would stop sending
us packets, but our Support folks should be well aware of those.  It had to
do with changes in NW Server continuations that were taking place in the
6.1.2 and 6.1.3 timeframe.  The 2.0 Exchange Module is based upon older
(5.7-era) client code that uses smaller chunks.

You might want to try running the recover command with diagnostics to see
where it stops.  Caveat: diagnostic output is undocumented, and output can
be counter-intuitive.

Hope that this helps,
_Scott

-----Original Message-----
From: Eichelberger, Jon [mailto:jon.eichelberger AT SAP DOT COM]
Sent: Tuesday, December 30, 2003 3:07 PM
To: NETWORKER AT LISTMAIL.TEMPLE DOT EDU
Subject: [Networker] Exchange Restores Hang


Hi, All.

Environment:
Networker Server (this is sssvr01) is a SUN V880 running Solaris 8,
Networker 6.1.3.
Storage Node (this sssb06b): SUN V880 running Solaris 8, Networker 6.1.3.
Client that is backed up (call this usphbx18): NT 4.0 SP6a running Networker
6.1.1, Exchange 5.5 SP4, Legato Exchange Module 2.0 Patch 4.
Client that is being restored to (call this usphbx1z): Same as backup client
except for Networker 6.1.3 (and later we tried updating the
Exchange Module to 2.0 Patch 6 on usphbx1z only with no change in the
result).
Jukebox: Scalar 1000 with AIT-3 drives.

Background: These restores had been working fine and recently the hangs
begun.  There was no event or change to which we can attribute the problem.

Problem: The redirected restore goes along fine for 42GB and then hangs.
The restore still needs to go for 8 more GB.  The network activity between
the storage node
sssb06b and usphbx1z goes down to an occasional packet of the form, as seem
via snoop on the storage node:
root@sssb06b>snoop usphbx1z
Using device /dev/ge (promiscuous mode)
     sssb06b -> usphbx1z TCP D=23279 S=9911     Ack=2126813101
Seq=3946588437 Len=1460 Win=32120
wait a minute of so...
     sssb06b -> usphbx1z TCP D=23279 S=9911     Ack=2126813101
Seq=3946588437 Len=1460 Win=32120
wait a minute or so...
     sssb06b -> usphbx1z TCP D=23279 S=9911     Ack=2126813101
Seq=3946588437 Len=1460 Win=32120
...

I tried restores of multiple backups from different exchange servers onto
usphbx1z.  They all hang at some point, not all at 42GB, but so far it has
always been at
42GB for any restore attempt of usphlx18's backup to usphlx1z.

According to the nsrxchrc.log on usphbx1z, this is what is going on:

Start time Fri Dec 26 11:39:11 2003
Computer Name: USPHLX1Z     User Name: Administrator
System Version: 4.0 Build 1381 Service Pack 6
ESEBCLI.DLL Version: 5.5.2653.11
ESEBCLI.DLL Comments: Service Pack 4
nsrxchrc.exe Version: 2.0.008 Build: 212
nsrxchrc.exe Comments: 2.0 Patch 4
Command line:
  nsrxchrc.exe -s sssvr01 -c usphbx18 -e -r -t Wed Dec 17 07:52:55 2003
 MSEXCH:IS
Active ANSI Code Page: 1252; OEM Code Page: 437
Non-cluster environment on node USPHLX1Z.
Options:
  Client: usphbx18
  Server: sssvr01
  Recover time: Wed Dec 17 07:52:55 2003
  Erase files: True
  Start services: True
Started PerfMon update thread (Id: 492)
Performance monitor update thread running
Recovering Microsoft Exchange Information Store ...
Recovering \\USPHLX1Z\E$\exchsrvr\MDBDATA\PRIV.EDB
(nothing more appears.  It's hung.)

Does anyone know of such a problem?  Our Legato support provider has not
come up with anything.

Thanks.

Jon

--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=

--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=

<Prev in Thread] Current Thread [Next in Thread>
  • Re: [Networker] Exchange Restores Hang, Eichelberger, Jon <=