SANsymphony-V v10 - Hands-on experience part II - Optimized Pool Recovery

Today I want to spent a bit time on the new features of SSY-V 10. These are:

  • Support for 40/56 GbE
  • Support for additional 16Gb QLogic FC HBAs
  • Support for Windows Server 2012 NIC teaming for iSCSI targets
  • Poller improvements
  • 32 nodes in a cluster
  • Virtual SAN
  • GUI improvements (system health and allocation maps)
  • Smart Deployment wizard
  • Pool space reservation for new writes
  • Improved vDisk rebalancing
  • Optimized Pool Recovery
  • Write-aware Tiering

 

The most important things for me to test are marked as bold in the list above. So let's start with the absolut most interesting feature, the Optimized Pool Recovery.

Optimized Pool Recovery means that if you lose a physical disk in a pool you don't have to rebuild the whole pool anymore but only the REALLY lost blocks on the missing disk will be rebuilt from the partner DCS.

Let's start a test. I have two DCS with each DCS attached to a MSA2000. Each MSA presents 11 disks to the corresponding DCS.All disks are configured within a single pool. Data already exists on the pool within 4 mirrored vdisks.

The additional "HP LOGICAL" disk shown in the pool is a local SSD used for AT.

Okay, as you can see all vDisks are up-to-date. I will now fake a hard disk error by unpresenting one of the physical disks. This will set the physical disk to be missing and will the pool to go offline.

As expected, the pool is now offline because of the missing disk. In SSY-V < v10 you now have to delete the whole pool and recreate it.

With version 10 you can purge this missing disk and all needed SAUs are copied from the partner DCS.

Purging the disk itself is quite easy. Simply right-click on the missing disk and select "Purge disk".

In this simple setup no further preparations have to be done so I'm immediately fine to purge the disk

As the vDisk VMFS001 is the only disk where changes happened during the last few minutes you see in the Step 2 of the wizard only this disk.

Step 2 also shows you which percentage of the data can be recovered. Luckily in our setup this will be 100%

After click on "Purge" the vDisks will rebuild but only in log recovery mode and not full recovery. In the status tab of the vDisk you can see that from the 660GB of data the vDisk contains, approx. 80GB were stored on the now missing disk and will re rebuilt from the second node.

Perfect! Without the new feature I would be supposed to replicate 1TB of data. Now it's shrinked to 80GB and with raising capacities and usage levels on each physical disk the benefit will be even bigger.

By the way, representing the purged disk to the DCS will NOT break any data consistency. The formerly used disk will show up as new disk under "Physical disks" and can be added to the same or a new pool.

We will use this in part III of this serie as we will test the new rebalance features.

 

Leave your comments

Post comment as a guest

0
Your comments are subjected to administrator's moderation.
  • No comments found
Powered by Komento
joomla templatesfree joomla templatestemplate joomla
2017  v-strange.de   globbers joomla template