Storage Boards Cleanup
To make it easier to find information about HPE Storage products and solutions, we are doing spring cleaning. This includes consolidation of some older boards, and a simpler structure that more accurately reflects how people use HPE Storage.
HPE StoreVirtual Storage / LeftHand
cancel
Showing results for 
Search instead for 
Did you mean: 

HP LeftHand Performance Issue? Or Not?

Peter J West
Frequent Advisor

HP LeftHand Performance Issue? Or Not?

We've been running the HP Lefthand series for a number of years now and although burst copy performance is sometimes a little slow it's generally performed quite well and we've not seen that much in the way of problems.

 

A few months ago we had a requirement to add more storage and this resulted in us buying two more StoreVirtual 4530 nodes.  These can't co-exist with the old P4500 G2's unless all of them are upgraded to at least LeftHand v11.5 so naturally one of the first jobs we had to do was get all nodes running the same revision of software.

 

We did this and the new nodes were added to the cluster without any issues.

 

However, soon after this we started to hear reports of odd behaviour.  The first thing I heard was that the backup was now taking much longer to complete.

 

There weren't any obvious performance issues at that time but as things have progressed we've started to notice a few things of concern.


Firstly we're seeing messages like this in vCenter:

 

Device
naa.6000eb39608bef62000000000000258e
performance has deteriorated. I/O latency
increased from average value of 12827
microseconds to 2882077 microseconds.
warning
09/03/2015 05:32:15

As a result of these messages I started doing some checked on storage latency on the paths and the averages seemed to be perfectly acceptable to me: 7.08m/s for write and 6.4m/s for reads with maximum values of 31m/s and 21m/s respectively.

 

Other things i've checked/done:

 

1. No alarms are visible in the CMC

2. The performance monitor graph in the CMC also shows a relatively low number of IOPS.

3. Created an additonal LH volume and spread the VM's out more - we had a number of SQL Servers on the same volume so I was quite sure that could cause some issues.

 

I'm curious if we do really have a problem or not.

 

I'd really appreciate any assistance that can be given in troubleshooting the issue.

 

Thanks

 

3 REPLIES
vlho
Advisor

Re: HP LeftHand Performance Issue? Or Not?

Hi,

 

I have good experience with disable delayed ACK.

Look here:

http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1002598

and try...

David_Tocker
Regular Advisor

Re: HP LeftHand Performance Issue? Or Not?

One thing to do, if your nodes are original without the 10gb kit is upgrade the RAM in them.

We noticed ours slow down after the upgrade to 11.5 so at the time I upgraded them to 12GB ram - the performance did not increase a lot if at all, but the consistency was improved - less errors being logged in the failover cluster manager in our windows Hyper-V cluster.

 

Now that 12.0 is out, finally they recommend upgrading the ram... probably should have been the case with 11...

Regards.

David Tocker
GarethRUK
Occasional Contributor

Re: HP LeftHand Performance Issue? Or Not?

+1 for delayed ACK.

 

We have a large number of P4500 clusters and VSAs. We are seeing the EXACT same issue.

 

I'm in the process of trying to change all the ESX nodes to use delayed ack + plus some other configuration changes or move over to the MPM module.

 

The MPM module is the preferred option for us as the installation can be automated using VUM.

 

 

Regards,

 

 

Gareth