Good morning Stephanie,
We were backing up Exchange 2007 just fine but when they moved to Exchange
2010 and we switched to NMM 2.4 we started having issues.
Matt Powell
mtpowel AT clemson DOT edu
864-656-0589 (Office)
864-247-2823 (Cell)
Storage Administrator
Clemson Computing and Information Technology
On Jan 22, 2013, at 10:57 AM, Stephanie Finnegan <sfinnega AT aip DOT org>
wrote:
> The timing does seem to point to an external event, and based on the fact
> that Excange is the only failure at that timepoint, i would expect to find
> the trigger within the Exchange environment. Have you looked at the system
> event logs on the Exchange server(s)?
>
> Also, did these backups work in a previous version of NW, or is this your
> first go at backing these up with Networker?
>
>
> Stephanie Finnegan
> American Institute of Physics
>
> On Jan 22, 2013, at 10:50 AM, "Matthew Powell" <mtpowel AT CLEMSON DOT EDU>
> wrote:
>
> Thanks Lars, for the suggestion but we have inactivity timeout already at 0.
> The tcp keep alive parameters have been set on both networker host and a one
> of the exchange servers to test with and it did not help. Still died but the
> funny thing is that the server never thinks the job fails until I kill it. It
> is still just sitting there expecting information from Exchange servers.
> Name resolution works on all systems. That was first think EMC wanted us to
> check. We have had this case with EMC going now for about a month with no
> production exchange backups.
>
> Matt Powell
> mtpowel AT clemson DOT edu
> 864-656-0589 (Office)
> 864-247-2823 (Cell)
> Storage Administrator
> Clemson Computing and Information Technology
>
> On Jan 22, 2013, at 10:34 AM, Lars Törnblom <lars.tornblom AT PROACT DOT SE>
> wrote:
>
>> Hello,
>>
>> One thing you can try is to set the inactivity timeout on the group to "0" =
>> disabled
>> if it helps you can test with a high value like 180.
>>
>> the TCP_KEEP Alive parameters have you tuned it on the backup server and the
>> exchange nodes.
>>
>> another thing to check is the name resolution within the exchange cluster.
>> the easiest way is to add all nodes and the DAG node to the hosts file on
>> all exchange nodes.
>>
>> br //Lars
>>
>>
>>
>> On 22 jan 2013, at 16:19, Tony Albers <Tony.Albers AT proact-externt DOT
>> dk<mailto:Tony.Albers AT proact-externt DOT dk>> wrote:
>>
>> Is the behaviour the same on all exchange servers? If so, could there be
>> firewalls/routers in between trying to shape the traffic, maybe dropping ACK
>> packets?
>>
>> Are all your exchange servers in the same savegroup in NW? Have you tried
>> splitting them up in several groups with different start times?
>>
>> How about cron on the backup server? Anything in there that might cause this?
>>
>> The fact that this happens at exactly 10 PM still points to something that's
>> scheduled somewhere.
>>
>> When the backups die, is there anything interesting in the nmm.raw file on
>> the client?
>>
>> Have you enabled savegroup logging on the networker server? Anything in the
>> log file at around 10 PM?
>>
>> Have you tried running savegrp -vvvc clientname savegroupname on the CLI on
>> the NW server?
>>
>> I have to go now, but I'll check my mail again later and maybe I've got some
>> more suggestions.
>>
>> /tony
>>
>>
>>
>> -----Oprindelig meddelelse-----
>> Fra: Matthew Powell [mailto:mtpowel AT clemson DOT edu<http://clemson.edu>]
>> Sendt: 22. januar 2013 16:06
>> Til: Tony Albers
>> Cc: EMC NetWorker discussion
>> Emne: Re: Exchange 2010 with NMM 2.4 and Networker 8.0
>>
>> I do backups to disk sir. I have been getting speeds of about 115 MB/s for 5
>> save streams from it. Yes it dies at 10:00 pm. There are no scheduled jobs
>> on exchange server that the group has found. If I just tell it to do 3
>> Information stores or less it works but that would take forever since we
>> have 6 servers with 10 stores each on them. Thanks for your help sir. I
>> appreciate it.
>>
>>
>>
>> Matt Powell
>> mtpowel AT clemson DOT edu<mailto:mtpowel AT clemson DOT edu>
>> 864-656-0589 (Office)
>> 864-247-2823 (Cell)
>> Storage Administrator
>> Clemson Computing and Information Technology
>>
>> On Jan 22, 2013, at 9:54 AM, Tony Albers <Tony.Albers AT proact DOT dk>
>> wrote:
>>
>> Exchange backup with NMM can be a real pain. But it usually works fine once
>> you get all the kinks straightened out.
>>
>> So, if the job starts at 8 PM, will it die at 10 PM that same day? That is,
>> 2 hrs later?
>>
>> The fact that it happens at that exact time would have me looking for
>> scheduled jobs or something similar that might affect the backups.
>>
>> How do you do the backups(to tape or disk or..), and at what speed?
>>
>> /tony
>>
>> -----Oprindelig meddelelse-----
>> Fra: EMC NetWorker discussion [mailto:NETWORKER AT LISTSERV.TEMPLE DOT EDU]
>> På vegne af Matthew Powell
>> Sendt: 22. januar 2013 15:48
>> Til: NETWORKER AT LISTSERV.TEMPLE DOT EDU
>> Emne: [Networker] Exchange 2010 with NMM 2.4 and Networker 8.0
>>
>> We are using Networker 8.0 server on Solaris 10 and we are having the worst
>> issues trying to back up Exchange 2010. I have an open case with EMC and
>> they are at a loss. I am just curious if anyone on this thread used Legato
>> to back up Exchange 2010 and if you are having any issues with it. Our issue
>> seems to stem from a timeout issue but we have set the TCP settings that EMC
>> has recommended but it has not helped. It is like the job makes the snapshot
>> and then starts to back it up and then just times out during the process and
>> it is always at 10:00 pm at night. It doesn't matter when the job started,
>> before 10:00 pm or after it will bomb by 10:00 pm the next day. Thanks for
>> your time and I hope everyone has a wonderful day.
>>
>>
>>
>> Matt Powell
>> mtpowel AT clemson DOT edu
>> 864-656-0589 (Office)
>> 864-247-2823 (Cell)
>> Storage Administrator
>> Clemson Computing and Information Technology
|
