2011-07-18 04:28 AM
We use netapp with vmware and have SRM for failover at DR.
Have just come across an issue where the FAS2040 we use for DR fails running SRM test plans because it has reached a limit on the number of volumes.
This is beacuse for a test SRM creates a flexclone of each of the volumes involved and we have reached such a point now that along with the mirrored production volumes when doing an SRM test it adds up to more than 200.
Does anyone know if there is a way around this problem as it pretty much makes SRM useless for DR testing.
2011-07-18 04:44 AM
in the technical overview of your fas2040, the limit for flex volumes is 200.
why you need such a big pool of volumens in your vmware environment?
have you create for each vm one volumen?
2011-07-18 05:00 AM
A large number of the servers have FC attached RDM's, there is a total of around 100 volumes in production they are being mirrored to DR, we are fine if we need to do a full dr test but ans the SRM dr test creates flexvols we cannot do a full dr test.
Having this fixed limit on the number of volumes seems a little mean especially when it takes flexvols into consideration in the count of volumes.
After doing some reading it seems that up to 500 is availiable on other models (we have 3140's in Propd) so it is possible to have more volumes, there must be a way to override this limit whist doing a test.
2011-07-18 07:02 AM
I share your pain here. I dont think you can override these settings.
One configuration that some customers use is that they include swap and paging files in their DR/backups.
I dont supposed you ve done that by any chance? if you have you could try to exclude them from your
DR set and see if that takes the number of volumes down enough for you to test DR.
Swap and paging files contain transient data and are typically not needed in DR circumstances.
Just a thought.
2011-07-19 01:55 AM
It seems ridiculous, you follow best pratices of 1 lun per volume then you get hit with a limit on the number of volumes.
We have a large number of sql servers and follow all the best pratices which say you should break out databases according to the snapshot schedules, pay out for all the snap licences, then netapp stuff you with a limit thats clearly set in software.
Is this worth raising with netapp i get an answer as to why this limit is there? or am I likely just get referred to the tech specs.
2011-07-19 04:46 AM
If you have VSC2.1 and SnapDrive 6.3 you could "move" some of the SQL servers to VMDK based disks, which "live" inside the datastore.
Maybe this could also help to bring the numebr of volumes down a bit.
2011-07-19 07:16 AM
The flexvol limit is in place due to resource limitations. That limit is better than it used to be, it used to be 200 for all systems with flexvol and 100 for tradvols.
Note: A high number of traditional and FlexVol volumes can affect takeover and giveback times if your system is part of an active/active configuration.