Forum Discussion

Rob_dos_Ramos's avatar
Rob_dos_Ramos
Level 6
13 years ago

Exchange 2010 Journaling PendingArchive Part

Hi All We have enabled journaling for a customer and we are seeing a large number of PendingArchive Part messages in the outlook profile. EV server is 9.0.3 and it is on a windows server 2008 R2 ...
Enterprise Vault
Information Governance
Performance
Troubleshooting
  • Prone2Typos's avatar
    Prone2Typos
    13 years ago

    So by teh design this is somewhat expected behavior. You need to note the time and date, if the items ever clear, clear slowly, or what the deal is in it. I have not seen anything implying that ... althogh I did review. Keep in mind that cycling the sercives will restart the review process and this is one of those issues that you have to step away from for a while to see it have a positive response.

     

    I have seen Exchange issus causing an issue, I have seen EV communication issues be related, and most recently... cached credential on the EV server relaing to the Journal performance. If you were haivng an issue communicating with the Exch Server, you should also see that manifest in the Archive task and it doing less than expected.

     

    The concurrent connections depends on a lot of env and sizing things... so I would not say there is a best practice outside of default. Keeping it as low as possible has typically been a cheap way to get out of fixing an issue (IMHO) and I would say is less than ideal. This is something that needs to be tuned, monitored, adjusted, and retuned to best suit your systems needs.

     

    Really to have an idea as to what is happening.. A DTrace is needed against the Journal task for ... I would say 15 minutes. You can use Pastepin or attache the log... please dont copy the contents of the trace into the forums... that drives me nuts personally.

     

    In a rare condition, I have seen some third party product (like antivirus on the Exchange server) modify the time stamps on these messages... which in turn causes them to never archive.

     

    I hope this helps. If you have had this resolved (as it has been a while since it was updated) please update the posting with teh resolution or mark the person that helped you as resolving it so that those out here to help can direct their attention to those who need it most.

     

    Many thanks.

AKL's avatar
AKL
Level 6
13 years ago

Just curious if issue is around the pending part archive items OR issue is with new items coming into journal mailboxes?

If issue is with pending part archive items - I will recommend to reset them to normal items using DocMessageClass utility.

We face such issues time to time in our environment, way I have set it up is I have extra journal databases with mailbox configured in EV as well but dismounted on Exchange side and disabled on EV side - I call them Helpforce Journal mailboxes.

When production journal mailboxes get overfull like yours - I bring those up quickly, point all journaling to new/idle mailboxes and start those tasks in EV - Because EV has issue when it has to go through 200000 items each time the thread restarts - and that causes delay as well. With new mailboxes - atleast EV can catch up with production journaling, backlog journaling clears out eventually.

Do you have DS Server key on EV pointing to a writable GC in site with NSPI limit set to fffffffff.? Also - Does your CAS array uses a load balancer? and if load balnacer, have you tried to point EV to particular CAS node using host file?