MSA 2312i drive replacement question

[Sirius]

HI ,

Just wated to check something before I proceed.

I have had a drive fail in my MSA, I have a drive ready to replace it with, is it best practice to replace and let it rebuild out of hours, or can I replace it during work hours?.

I wasnt sure if the rebuild process would affect the performance of the unit in any way.

thanks.

[HPSDMike]

Best practices would dictate that you should replace the drive as soon as you have a replacement available (hence the concept of hot spares). There is always a chance of a performance hit but that is typically preferable to complete data loss that may occur (depending on your configuration) if another driver were to fail in the meantime. 

[Sirius]

Ok thanks, I already moved the VM's off of that array, so its empy now.

Would the performance hit affect the entire enclosure?, or just the array it is working on?

[HPSDMike · Accepted solution]

In general, people replace failed drives in production systems all day every day so any potential performance hit is generally not considered significant. However, the answer to your question unfortunately is "it depends." In general, the performance of that vdisk may suffer slightly while the array is rebuilt. However, if the controllers performance of the array are already maxed out then the rebuild may tax those controllers even move contributing to a potential system wide performance problem. You could explicitly set a 'utility-priority' if you want to lower the priority of the rebuild as to minimize contention. See page 81 of  https://support.hpe.com/hpsc/doc/public/display?docId=emr_na-c01755995 

[Torsten.]

After replacing the disk, make sure to make it a dedicated or global spare, otherwise it is not a spare and won't be used at all.

[Sirius]

Thanks all,

I replaced the drive and made it a global spare, it took around 5 hours to fully reconstruct.