Jon_G1
1 Copper

Length of time for recovery of RAID group following disk failure

Jump to solution

Hi,

I am composing a DR document and need to know what is the expected  length of time for the recovery of RAID group following a disk failure in a Clariion CX-4 storage array for the following disk types:

  • RAID 1 - SATA 7K, 1TB
  • RAID 5 - SATA 7K, 1TB
  • RAID 1 - FC 15K, 450GB
  • RAID 5 - FC 15K, 450GB
  • RAID 1 - SSD, 130GB

I have so far been unable to find any literature that gives me any indication so your help will be greatly appriciated.

Thanks

Labels (2)
0 Kudos
6 Replies
Highlighted
AnkitMehta
4 Germanium

Re: Length of time for recovery of RAID group following disk failure

Jump to solution

Hi Jon,

As far as I am aware there is no such document and one can not provide precise time period for the same.

Once the drive fails it goes through below states:

Failure - Rebuilding to Hotspare - Replacement - Equalizing.

The the recovery of RAID group following a disk failure depends on many factors right from its hardware...Drive Type, Size, Spindle Speed etc.  I understand you have already mentioned the part number which provides those details. However, there's more factors just than that...type of the data, applications, I/O's happening on the RAID Group etc. which varies on disks on different Raid Groups.

For example: 1TB EFD/SSD Drive failure with Normal Data blocks takes minimum of ~5 Hours.

Jon_G1
1 Copper

Re: Length of time for recovery of RAID group following disk failure

Jump to solution

Thank you for your quick response.

Any idea on the predicted minimums for the other drives mentioned? This does not have to be 100% accurate, just an indication of what to expect.

0 Kudos
AnkitMehta
4 Germanium

Re: Length of time for recovery of RAID group following disk failure

Jump to solution

Hi Jon,

You're welcome. I am still working on this Question. Please, allow me some time.

It would be difficult to give you the exact time but I'll try my best to give you the approximate minimum time.

Moreover, it will be based on Drive Type and Size not for RAID Group + Drive Type + Drive Size.

0 Kudos
kelleg
5 Rhenium

Re: Length of time for recovery of RAID group following disk failure

Jump to solution

The CX4, please see page 89 of the Best Practice guide - in the section called "Rebuilds"

EMC CLARiiON Performance and Availability Release 30 Firmware Update Applied Best Practices.pdf

http://powerlink.emc.com/km/live1/en_US/Offering_Technical/White_Paper/h5773-clariion-best-practices...

For VNX see page 68 - Drive Rebuilds section

EMC Unified Storage Best Practices for Performance and Availability - Common Platform and Block Storage 31.5 — Applied Best Practices.pdf

http://powerlink.emc.com/km/live1/en_US/Offering_Basics/White_Paper/h8268_VNX_Block_best_practices.p...

Glen

tkjoffs
2 Iron

Re: Length of time for recovery of RAID group following disk failure

Jump to solution

One would think that this would depend on the actual number of LUNs and the data on the Raid Group.  Since I believe the rebuild and equalization of a Raid Group only rebuilds the actual used capacity, your minimums for a Raid Group rebuild and equalization would be less than 5 minutes if the Raid Group was newly created and had no data in it.  It could be upwards of 10 hours if you are rebuilding a Huge Raid Group fully populated.  This also depends on the Rebuold Priority (use Low/Medium if you don't want to be fired the next day).

0 Kudos

Re: Length of time for recovery of RAID group following disk failure

Jump to solution

This is true.  FLARE rebuilds LUNs, not RAID Groups, so if you have a 2TB RAID Group with just a single 50GB LUN, FLARE only rebuilds the data for the 50GB LUN.  The best practices guides for CLARiiON and VNX provide average MB/sec rebuild rates which depend on the rebuild priority and system load..

EFD disks:

Priority   -   MB/sec rebuild

Low      -    2

Medium -    6

High      -    13

ASAP   -    160

15K Disks:

Priority   -   MB/sec rebuild

Low      -    2

Medium -    6

High      -    13

ASAP   -    70-80

As you can see, there is a very large increase in rebuild rate for the ASAP priority in exchange for a potential performance impact.

0 Kudos