08-28-2011 07:32 PM
Hi
I am having 1 EV 9.0.2 server and Approximately 13 Mailbox arhving task out of which 4 our for my Exchange 2010 server and each task for exchange 2010 has 4 mailbox stores in its Target tab.
Now we are facing issues that the schedule archive is not working, when tried to archive mails manually it is working fine.
When looking at queues i see messages stuck in A3, A5 and A6 queue. Also i have kept the Archving task to run for 24 hours on site level.
Any inputs on where to look ahead on this.
Solved! Go to Solution.
08-31-2011 08:48 PM
Hi,
So you have no MAPI related errors and you have given a snippet of DTRACE which tells me that there are thousands of items on your storage archive MSMQ and it is throttling it. Take a look at the following technical article:-
http://www.symantec.com/docs/TECH35772
You can change these upper and lower limits via a registry key but the question really is why is it getting to this level in the first place? With that amount of tasks I may suggest that you are asking the server to do too much.
Do you have any details on how much you are expecting to be archived each hour within your archiving schedule over the 13 exchange servers?
I would suggest that this is a badly designed/configured system but unless a full trace is analysed the see why these items are building up on the storage queue it is impossible to say if any tweaks could be made to improve the situation or more hardware needs to be thrown at it. As explained previously that many tasks on one server would only work if we are looking at archiving servers at different times and if there is not a substantial amount to archive each scheduled run.
Cheers...Paul.
08-28-2011 08:34 PM
Without seeing any evidence it is difficult to state what your issue may be. Straight away though the fact that you have 13 exchange mailbox tasks running on the same EV server points to MAPI resource issues. 13 tasks is too many tasks as there are finite MAPI resources available and I am assuming that they are running at the same time.
Are there any event errors in the EV event log which indicate MAPI resource issues?
If you only have a subset of these tasks running then is the system archiving?
How are you determining that nothing is being archived?
Really the only way to see what is happening is firstly by running a dtrace against the archivetask process and analyzing it. Maybe then further processes may need to be traced on the storage side.
The A3 queue is used when you do a run now manually against a mailbox so that queue can be purged, the A5 queue is used for scheduled archiving so this shows that mailboxes have been put onto the queue for processing and the A6 queue is used for updating moved items.
08-28-2011 09:29 PM
08-28-2011 10:24 PM
@Paul answers to your questions
Are there any event errors in the EV event log which indicate MAPI resource issues?
No i havent seen any errors regarding MAPI
If you only have a subset of these tasks running then is the system archiving?
Not sure about this
How are you determining that nothing is being archived?
I tried to run below query on sql server and i can see the count going on very slowly also my mailbox is also not been archived since which should have been according to policy.
select count(*) from journalarchive
@Jesus.
I can see 3305 and 3310 errors but that is intermittent. Also i have re-scheduled the site settings with 15 mins gap interval.
Outlook 2007 Hotfix has also been applied on EV server.
08-28-2011 11:00 PM
is there anything specific i can have a look at dtrace logs
08-30-2011 04:14 AM
Dtrace shows this
6 16:53:56.296 [740] (ArchiveTask) <10408> EV:H {CArchivingAgent::ArchiveMarkedItems()} (Exit) Status: [While processing mailbox %1 the Storage Service reached its lower MSMQ threshold. The Archiving/Journal Task will pause to let it recover. (0xc0040b31)]
36 16:53:56.359 [740] (ArchiveTask) <10408> EV:M AgentMessageDispenser: 9 of 9 Agent threads is pausing
30 16:53:56.296 [740] (ArchiveTask) <10408> EV:M AgentMessageDispenser:rocessNextMessage - AGENTS_E_DISPABORTITEM, agent will sleep and abort the the current MSMQ Transation
1322 16:54:16.030 [1232] (ArchiveTask) <12044> EV:M CCAAgentHelper::ArchiveMarkedItems - Error [0xC0040B3
And they want me to reinstall the binaries on the EV server.
08-30-2011 05:40 AM
That does sound stupid, reinstalling the binaries won't get you far
Just as a matter of interest, how big is the MSMQ storage directory at the moment?
08-30-2011 07:54 PM
MSMQ storage size is set to 8 Gb and the size is of MSMQ is approx 4 GB
08-30-2011 08:39 PM
Was there deviation from EV Best Practice Registry Settings during installation, or since?
08-30-2011 09:10 PM
Nothing as such has been done on the server in terms of registry, can you please let me know any specific registry which you want me to check.
08-31-2011 08:48 PM
Hi,
So you have no MAPI related errors and you have given a snippet of DTRACE which tells me that there are thousands of items on your storage archive MSMQ and it is throttling it. Take a look at the following technical article:-
http://www.symantec.com/docs/TECH35772
You can change these upper and lower limits via a registry key but the question really is why is it getting to this level in the first place? With that amount of tasks I may suggest that you are asking the server to do too much.
Do you have any details on how much you are expecting to be archived each hour within your archiving schedule over the 13 exchange servers?
I would suggest that this is a badly designed/configured system but unless a full trace is analysed the see why these items are building up on the storage queue it is impossible to say if any tweaks could be made to improve the situation or more hardware needs to be thrown at it. As explained previously that many tasks on one server would only work if we are looking at archiving servers at different times and if there is not a substantial amount to archive each scheduled run.
Cheers...Paul.