Forum Discussion

Nate_D1's avatar
Nate_D1
Level 6
13 years ago

EV Backup questions

Hi All,

 

I have been struggling with backups of our EV environment for the past few months, and thought I would turn to the community for a bit of advice.

We currently run EV10.1 and NetBackup 7.5.0.3 backing up with the 'Enterprise-Vault' policy type.

 

I am running into two seperate issues, first is that EV does not seem to be deleting all safety copies. We have mail setup to archive after 30 days, delete shortcut after 90.  I have a chunk of time from 11/22/12 to 12/5/12 where the safety copies did not delete, and it has the icon of the disk with the clock on it (rather than the 2 squares). But everything from 12/5/12 to 12/12/12 is archiving and deleting copies as normally.

Is there something I can do to force the shortcuts to process during that time period?

 

Second issue is that my log files on the SQL server are growing really large (4-5x the size of the database), I think this is due to only a certain type of backup (I think cumulative incremental?) will truncate the logs. I remember talking about this at EV class, but I dont remember what the specifics were. I had been running Full backups for the past couple weeks because I thought that would clear the safety copies every night, but it does not seem to be truncating the logs.

 

So I would like to hear what some others do for their backups, to handle truncation and safety copies frequently. I plan to dig through the 'best_practice_for_ev_backups.pdf' this week as well to review this. 

 

Thank you for any help you can provide!

Backing Up
Best Practice
Enterprise Vault
Information Governance
Troubleshooting
  • JesusWept3's avatar
    JesusWept3
    13 years ago

    For the pending items, the best bet would probably be to set the policy (either journal or mailbox) to set a Pending Shortcut Timeout of 5-10 days and then run against the mailbox

    Because most likely if you are sure that the backups genuinely are working and EV isn't reporting a backlog of items that have not been backed up, then most likely what happened is that it never made to storage in the first place.

    I.e. it was placed in pending mode, but failed to archive correctly.
    Another thing it could be is if the item was in pending status and awaiting backup and then the pending item was moved from its original location elsewhere, then the item would remain in pending as the Task that would change the item could not find the pending item altogether

    Check this article, particular Solution #2
    http://www.symantec.com/business/support/index?page=content&id=TECH35618

    As for the transaction logs, it is normal for the log to grow and grow and become larger than the database as it will simply log all insert, updates, deletes etc.

    This may help with your issue?
    http://www.symantec.com/business/support/index?page=content&id=TECH158100

  • JesusWept3's avatar
    JesusWept3
    Level 6
    13 years ago

    For the pending items, the best bet would probably be to set the policy (either journal or mailbox) to set a Pending Shortcut Timeout of 5-10 days and then run against the mailbox

    Because most likely if you are sure that the backups genuinely are working and EV isn't reporting a backlog of items that have not been backed up, then most likely what happened is that it never made to storage in the first place.

    I.e. it was placed in pending mode, but failed to archive correctly.
    Another thing it could be is if the item was in pending status and awaiting backup and then the pending item was moved from its original location elsewhere, then the item would remain in pending as the Task that would change the item could not find the pending item altogether

    Check this article, particular Solution #2
    http://www.symantec.com/business/support/index?page=content&id=TECH35618

    As for the transaction logs, it is normal for the log to grow and grow and become larger than the database as it will simply log all insert, updates, deletes etc.

    This may help with your issue?
    http://www.symantec.com/business/support/index?page=content&id=TECH158100

  • Nate_D1's avatar
    Nate_D1
    Level 6
    13 years ago

    I do expect the logs to grow, but I dont expect it to be larger than the DB. They have never been larger than the db themselves, often being *around* a 1:1 ratio.  I am guessing because I switched to 'nightly fulls' instead of 'nightly incrementals, weekly fulls' that i stopped truncating when I stopped the incrementals.

     

    I will followup with the links to fix the archive pending emails, thanks!

     

     

  • Nate_D1's avatar
    Nate_D1
    Level 6
    12 years ago

    To follow up on this, I am still working on issue #1, but issue #2 is resolved.

    I beleive the issue is that a Full backup does not truncate the log, only a Differential backup will truncate the log.

    Below was my fix, im not a SQL expert this is just what works for me :)

    I switched back to differential incrementals during the week, and a full on weekends. After doing that I got on the SQL server and ran

    DBCC SQLPERF (LOGSPACE)

    Which returns the size of all of the logs, and the % of used space in the log file. They were all still packed 96%+ and 4-5x the size of the database.  I ran each database through a manual backup/shrink/ process that I use on some other databases

    USE (database)
    Go
    DBCC SHRINKFILE (Database_LogName, 30);
    GO
    backup log Database to disk='Y:\file.trn'
    go
    DBCC SHRINKFILE (Database_LogName, 30);
    GO
     

    I ran this process on each of my databases, then ran a full backup with NetBackup (it will error if you try and do another incremental, saying the last full was not done with netbackup and it needs to re-do it)

    Everything is good in that arena now :)