MSA 2040 Version Checks Related to General Health and specifically VDRAIN

[Lekhpal]

Team. We came to know that sometime VDRAIN process can go on for months. In one specific case, after MSA Firmware was upgraded, the problem was solved.  We've a MSA2040. So, we want to know what is our MSA Firmware version. Can it be self upgraded to latest version. What are the steps. Does our current MSA Firmware is affected with this bug of VDRAIN  going on for months(never seems to be finishing). Can we upload the logs to https://msa.ext.hpe.com and get the recommendations on which MSA Firmware to update to, and also the steps to do that ourself etc. And, finally, from where we can download the latest binaries for the MSA 2040.  BR

[JonPaul]

@Lekhpal 
VDRAIN can take a long time no matter the firmware version.  You can find all available MSA firmware at this site:  https://www.hpe.com/storage/MSAFirmware
This will have direct links to HPE download sites.  

Using the MSA HealthCheck site https://www.hpe.com/storage/MSAHealthCheck  (Points to same link you provided)  is a great step to seeing best practices for your array.  Best Practices include recommended firmware revisions.

[Lekhpal]

@JonPaul Thanks John for the great info.

Another question:  Our Pool B has a failed disk, and now we are going to entirely deleting this Pool B. Once Pool B is deleted, we want to re-create the Pool B, with a Global Hot-Spare. What are the steps/commands to accomplish that. Pool A is good, and has no disk failure there. But we have A and B as RAID6 configurations, and no Global Hot Spare anywhere. What are the steps for that. Note, in show advanced-settings , we do see that 'Dynamic Sparing' is disabled. So, still we can create a Global Spare or not ?  If not, then how to enable 'Dynamic Sparing. The goal is have one disk as a Global spare, and can pick , if any of the drive anywhere fails.  BR.

[JonPaul · Accepted solution]

@Lekhpal 
NOTE: DELETING A POOL DELETES THE DATA, please make all necessary checks that your data is safe and stored elsewhere.
The quickest way to delete data and the pool is from the CLI

CLI> delete pools B    ==> in IN210R004 firmware there is no confirmation prompt, be really careful with this command

A more cautious approach would be to:
- Unmap the volumes from the B controller 
    - One at a time and validate that your hosts still have access to all the data they should
- Delete the volume that was unmapped
When all volumes have been removed
- Remove disk-groups  
   I think you only have 1 disk-group but with a pool with more than one disk-groups when one is removed any data on that disk-group will drain to the others. With no data on the disk-groups (Volumes removed) it should not take long. If it takes more than a couple minutes there may be a volume or 2 you missed.
- When you remove the last disk-group the pool will also disappear
   Any data on the disk-group will be gone
That is a cautious approach, if anything doesn't seem right you can double check to mae sure you are removing what you think you are removing.

The MSA HealthCheck (www.hpe.com/storage/MSAhealthCheck) recommends enabling Dynamic Sparing.  Dynamic Sparing will allow the system to utilize any AVAIL drive that matches the type needed to be used as a spare.  If Dynamic Sparing is disabled a drive must be set as a SPARE in order to be used to rebuild a disk-group/RAID set.  It's also a good idea to set SPAREs when that is the intentional use of a drive otherwise you are allowing the system to choose which AVAIL drive should be used as a spare,  i.e. system could use a 2.4TB 10k drive instead of the 1.2TB 10k to rebuild.
With the MSA Virtual Storage process all SPAREs are GLOBAL, meaning that they will be used for the first disk-group that is degraded.  If you are using Linear Storage then SPAREs can be limited a specific disk-group but the SPARE can also be chosen as a GLOBAL. See the CLI guide.
When creating an MSA 2040 RAID disk-group you should pay attention to the Power of Two rule.  Where the data drives are a power of 2 (2, 4, in the RAID set.  See MSA 2040 Best Practices guide (same caution as in 2050 and 2060).  If your data workload uses Sequential WRITEs, the performance can/will be limited.