cancel
Showing results for 
Search instead for 
Did you mean: 

Free Up Index Space - Journaling Archive & Retention.

AKL
Level 6
Certified

Hello All

This forum is second part of below showm forum. Kindly open same for quick background:

https://www-secure.symantec.com/connect/forums/change-retention-category-historical-archives

We're running Evault 9.0.1.1073 and as discussed on above forum, we cleared out nearly 8000K+ from journaling archive. However, we do not see corresponding increase in index space. We see the increase but its very minimum.

Just curious if any manual effort is required after expiring items so they're cleared from indexes as well? I am hesitant to do rebuild of indexes on journal archives without proper confirmation that it is mandatory to do so...

1 ACCEPTED SOLUTION

Accepted Solutions

JesusWept3
Level 6
Partner Accredited Certified

So take a look at this query and give it a run

SELECT A.ArchiveName, (IRP.IndexRootPath + '\'+ IV.FolderName) AS Folder, IV.FirstItemSequenceNumber AS FirstISN, IV.HighestItemSequenceNumber AS LastISN, IV.IndexedItems, IV.Rebuilding, IV.Failed, IV.FailedItems
FROM IndexRootPathEntry IRP,
IndexVolume IV,
Archive A,
Root R
WHERE A.RootIdentity = R.Rootidentity
AND R.RootIdentity = IV.RootIdentity
AND IV.IndexRootPathEntryId = IRP.IndexRootPathEntryId
AND A.ArchiveName = 'your Archive Name'

This should give you the full physical paths of the folders, the older the indexes are the more it should have been impacted by expiry, so pick one, go to the folder, figure out how big the folder is...then create a new file called compact.task, don't call it anything else, then when you search or update it will see the compact.task and then, well, compact it.

Then take a look again at how big the directory is

https://www.linkedin.com/in/alex-allen-turl-07370146

View solution in original post

6 REPLIES 6

JesusWept3
Level 6
Partner Accredited Certified
When items are expired and deleted they go to the JournalDelete table and have a "IndexCommited" flag on each record If its set to 1 it means the index no longer holds the item, if it's 0 it means index is still awaiting to delete the item Do a quick count on the JD table where indexcomitted=0 and see what you get Other than that, how long ago did this occur? You may need to do a compaction, try creating a compact.task file in one of the index volume directories, do a search on that volume and look and see if it gets any smaller
https://www.linkedin.com/in/alex-allen-turl-07370146

AKL
Level 6
Certified

Just quick confirmation, the table is showing values as true or false.

If i go by binary - 0 = False and 1 = True, same holds in this scenario as well right?

JesusWept3
Level 6
Partner Accredited Certified
Yup, the columns a bit/Boolean so it will be 0/False and 1/True
https://www.linkedin.com/in/alex-allen-turl-07370146

AKL
Level 6
Certified

Thanks for confirmation Alex.

Just another quick question (I feel like noob while asking it). How do i determine which folder I need to create compact.task file in?

JesusWept3
Level 6
Partner Accredited Certified

So take a look at this query and give it a run

SELECT A.ArchiveName, (IRP.IndexRootPath + '\'+ IV.FolderName) AS Folder, IV.FirstItemSequenceNumber AS FirstISN, IV.HighestItemSequenceNumber AS LastISN, IV.IndexedItems, IV.Rebuilding, IV.Failed, IV.FailedItems
FROM IndexRootPathEntry IRP,
IndexVolume IV,
Archive A,
Root R
WHERE A.RootIdentity = R.Rootidentity
AND R.RootIdentity = IV.RootIdentity
AND IV.IndexRootPathEntryId = IRP.IndexRootPathEntryId
AND A.ArchiveName = 'your Archive Name'

This should give you the full physical paths of the folders, the older the indexes are the more it should have been impacted by expiry, so pick one, go to the folder, figure out how big the folder is...then create a new file called compact.task, don't call it anything else, then when you search or update it will see the compact.task and then, well, compact it.

Then take a look again at how big the directory is

https://www.linkedin.com/in/alex-allen-turl-07370146

AKL
Level 6
Certified

Thanks a lot Alex - This was helpful.

As an update, I ran the query and found 104 records which had IndexCommited = False and 403 Index locations using the second query.

I created compact.task file in nearly 50 of them and ran search on journal archive. However still do not see significant increase.

The folders reported by query, were not large either compared to the new ones. So I am guessing it cleared out max drive space it already could?

I will try to run DTrace on indexing component as well to see what EV is doing there. Any other thoughts in this direction?