Discussion on: Using batching to do large operations without filling the transaction log http://itknowledgeexchange.techtarget.com/sql-server/using-batching-to-do-large-operations-without-filling-the-transaction-log/ Tue, 07 Oct 2008 09:13:37 +0000 http://wordpress.org/?v=wordpress-mu-1.2.1 By: Mrdenny http://itknowledgeexchange.techtarget.com/sql-server/using-batching-to-do-large-operations-without-filling-the-transaction-log/#comment-83 Mrdenny Tue, 19 Aug 2008 19:13:31 +0000 http://itknowledgeexchange.techtarget.com/sql-server/using-batching-to-do-large-operations-without-filling-the-transaction-log/#comment-83 No it's not. There's no way to batch index rebuild operations. By changing the recovery model from bulk-logged to simple you are breaking your transaction log series which will stop you from being able to restore your transaction logs. Instead you should trigger a transaction log backup as soon as your reindex job is complete. About all you can do would be to do log backups between each index rebuild command. You'll need to keep that amount of drive space reserved for your transaction log. No it’s not. There’s no way to batch index rebuild operations.

By changing the recovery model from bulk-logged to simple you are breaking your transaction log series which will stop you from being able to restore your transaction logs.

Instead you should trigger a transaction log backup as soon as your reindex job is complete.

About all you can do would be to do log backups between each index rebuild command. You’ll need to keep that amount of drive space reserved for your transaction log.

]]> By: YSLGuru http://itknowledgeexchange.techtarget.com/sql-server/using-batching-to-do-large-operations-without-filling-the-transaction-log/#comment-82 YSLGuru Tue, 19 Aug 2008 17:50:26 +0000 http://itknowledgeexchange.techtarget.com/sql-server/using-batching-to-do-large-operations-without-filling-the-transaction-log/#comment-82 Is it possible to apply this same concept of minimal impact to the Log file when performing maintenace tasks like index rebuilds? We have a 100GB DB with 2 very volatile tables which require Index Rebuilds weekly on a total of 6 indexes. The Log file has 60GB of space to work with and under any other DB activity the log file seldom gets to a fourth of that size. But when we rebuild these indexes, all 60GB are consumed. We have the RECOVERY MODEL set to Bluk-Logged, something we can't change due to requirements by Microsoft's DPM Software which we use for backups. Curently I'm changing the Reovery Model to SIMPLE then running the Rebulds and switching back to Bulk-Logged once the job is done. This occurs outside normal user hours so there's nothing else going on with the DB when the rebuilds occur. Even though this approach works, I'd love to finid a way to do this without changing the Recovery Model and without always having such a large amnount of drive space avaiable for the log file. Thanks Is it possible to apply this same concept of minimal impact to the Log file when performing maintenace tasks like index rebuilds? We have a 100GB DB with 2 very volatile tables which require Index Rebuilds weekly on a total of 6 indexes. The Log file has 60GB of space to work with and under any other DB activity the log file seldom gets to a fourth of that size. But when we rebuild these indexes, all 60GB are consumed.

We have the RECOVERY MODEL set to Bluk-Logged, something we can’t change due to requirements by Microsoft’s DPM Software which we use for backups. Curently I’m changing the Reovery Model to SIMPLE then running the Rebulds and switching back to Bulk-Logged once the job is done. This occurs outside normal user hours so there’s nothing else going on with the DB when the rebuilds occur. Even though this approach works, I’d love to finid a way to do this without changing the Recovery Model and without always having such a large amnount of drive space avaiable for the log file.

Thanks

]]>