HPE Storage Users Group
https://3parug.com/

Manually Releasing Spare Chunklets
https://3parug.com/viewtopic.php?f=18&t=3190
Page 1 of 1

Author:  thebossbaby [ Tue May 14, 2019 9:53 pm ]
Post subject:  Manually Releasing Spare Chunklets

Hello 3par admins,

I'm in a situation where I may need to release some spare chunklets for additional raw capacity. I understand that it is not recommended but this process will only be temporary.

From my understanding, spare chunklets are automatically set aside initially to be used for when a physical disk fails and logical disk raid set needs to be rebuilt.

Let's say for example, if I have 16 SSDs (1787 GB each) in a raid-5 setup and the system has reserved 2500 GB for spare chunklets, can I safely release 700 spare chunklets?

Author:  MammaGutt [ Wed May 15, 2019 12:15 am ]
Post subject:  Re: Manually Releasing Spare Chunklets

That depends on your definition of safe and how fast you would get replacements for broken drives.

Author:  markinnz [ Wed May 15, 2019 4:54 pm ]
Post subject:  Re: Manually Releasing Spare Chunklets

Not really I think ... you can modify the sparing algorithm to reduce the number of spare chunklets but I don't believe you can for "spare" chunklets to become avaiable for data use.

"showsys -param" will show the current sparing algorithm.
"setsys" will let you change it, and "admithw" will start doing the chunklet reshuffle.

3PARs are quite heavy for spare space, the ones I look after have 192 *3.84TB drives, and reserve 48TB of spare using the "default" algorithm (equates to 250GB per drive) ... seems a tad excessive to me :-)

Author:  Richard Siemers [ Thu May 16, 2019 1:07 am ]
Post subject:  Re: Manually Releasing Spare Chunklets

Looks like you know about the removespare command... your question is not "how", but more of "should I". The official HPe recommendation is raid6 with standard sparing.

Some businesses can tolerate downtime and handle restores after hours or weekends, while others lose money every minute the system is down and risk averse. What are you and the business comfortable with? It's your decision, and risk to own. Kick the question up the ladder along with your professional recommendation. Present the options, explain the risk, and outline the process to recover if the risk becomes an outage event. They can accept the risk, or upgrade capacity, or perhaps volunteer to delete some of those QA/TEST systems they asked for but never use :)

Some things I recommend you factor into your decision:

Check for any zero reclaim/compactcpg opportunity to get space back. You may need to manually run some commands on the host to zero out whitespace on older operating systems. I generally prioritize largest volumes with the highest used% first.

If you have any Fully Provisioned volumes, you may be able to reclaim space by converting them to thin.

Make sure you are not missing any patches/firmware that address SSD failure rates.

Your drive failure history. Is it one drive a week, or one drive a year?

If you have an ASM/Proactive Care Advanced, discuss this with them, verify part availability and gauge expectations for drive replacements. The longer the ETA for part, the longer you would be at risk of double disk failure.

Page 1 of 1 All times are UTC - 5 hours
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group
http://www.phpbb.com/