SIS Changelog Keeps Filling Up

[TMADOCTHOMAS]

For the last couple of days, the sis changelog on one particular volume that contains VMware datastores keeps filling up. We have several VMware datastore volumes and this is only happening on this one volume. It's one of the smaller ones at 250 GB. i re-run the volume efficiency start with -scan-old-data set to true to clear the error, but I can't figure out why it keeps occurring. This is an AFF system with inline dedupe/compression on, and scheduled dedupe which takes about 5 - 6 minutes. I did change the scheduled dedupe time to twice a day instead of once but that didn't clear the issue. Any ideas?

[colsen]

Hello,

Not sure how applicable this is to your situation, but we had a couple of Oracle log volumes that we ran into this same issue on.  The end result is that we weren't really getting much benefit from the scheduled dedupe and all that kept happening is the changelog would get full constantly.  We ended up just disabling scheduled dedupe on almost all our Oracle volumes, focusing on compression/compaction and then just letting inline do whatever good it could.  We actually improved our efficiency numbers a bit by going with that approach and avoided the changlog errors to boot.

Hope that helps,

Chris

[TMADOCTHOMAS]

Thanks colsen. I thought if I turned it off it would reach 100% sooner. I may try turning it off and see if it helps.