Storevirtual 3200 Latency Issue

[ThomasU21]

Ok, after a LOOOOng google run I found the FOM for sv3200...HP your Website is the worst...ever...It was good once...but since you separated into HP inc and HPE, its gets worse...from day to day...thanks a lot.

Now I am still searching for the FOM sv3200 13.6...I cannot install 13.5 and add it to the cluster. I cannot update it nor create a mgmgroup on the FOM to update it etc.

Any Help would be appreciated.

regards

Thomas

[mvbk]

Thomas,

you won't find a FOM for Storevirtual OS 13.6 for now. This is the feedback from HP:

Yesterday I have checked with L3 team and 13.6 ovf file for FOM is expected to be released on 27th Nov 2017. We can use that ovf file and build FOM and add to the existing array which is running 13.6.

So for now there is no way to add a new FOM to your 13.6 management group.

I don't need a FOM to present my volumes or prevent data loss if one controller fails, but what if a storage node goes down? I will lose 2 out of 4 managers, so thats why I'm using a FOM to have 3 out of 5 managers if a storage node fails. At least that's what I believe we should use? All my volumes are in NR10, so I asume the storage is smart enough to choose one controller from each storage node when setting up the NR10?! Unfortunately I haven't found a way to check this yet... I'm open for suggestions or advice if the FOM isn't necessary.

[LeeTse]

Hi ThomasU2 & vmbk,

what is your situation, i have upgrade to LFOS 13.6 and FOM 13.6 last week, but no performance improve, the performance is really worse, the array attached as ESXi server storage, copy files in windows guest machine running on ESXi, the speed is down to 800K/s,  and not exceed 60MB/s.

[peyrache]

Hi
New patches will fix this issue, will be available in next couple of days
JY

[ThomasU21]

Hi All

I just checked. I get around 50-60MB/s when I copy a huge file within an VM on an ESXI (iscsi attached sv3200)

I also found the FOM 13.6 to download on the HPE page.

Lets wait for the patches then.

Thomas

[Nor3344]

And then HPE announced EOL for the SV3200...

https://support.hpe.com/hpsc/doc/public/display?docId=emr_na-a00041272en_us&hprpt_id=HPGL_ALERTS_2000822&jumpid=em_alerts_us-us_Feb18_xbu_all_all_1462192_2000822_Storage_recommended__&DIMID=EMID_5F1E1D04110235971782C901259FE98E/

This smells like HPE is having to many ssues with the system that they just EOL it.... Support will be there for exsisting customers, but I do not take this as a good sign for the issues people seem to be having.

[referencepoint]

The icing on the cake for this whole issue I think. What an utter farce.

[marc5000]

We recently got an SV3200, the order went through during the EOL notification!

When the unit arrived it had a dead controller - we also found out (rather annoyingly) the unit wasn't fit for XenServer. 

Anyhow.. with the unit's second controller replaced,  we hit a few more snags. I was not eager to spend ages with HPE support (they were very good) - In the end, as it wasn't in production,  a complete flash of the firmware back to factory sorted the unit out. I started a new lab environment. 

Unit currently consists of 21x 900gb 10k SAS drives, RAID6 across 20 of them (one hot spare). We're going to continue testing before jumping on the Advanced Optimization SSD's - we have 2x, but realised we needed 4 (reseller never informed us). 

I'm running 10Gbps connections in ALB pairing on both controllers through a switch (awaiting second for proper multipath config). I have one Hyper-V node (Lenovo), with dual 10Gbps connections back to the switch. 

StoreVirtual OS is currently fully patched on version: 13.6.00.0241.0

Anyway - results form the raid6, as I expected.. not bad. Write performance maxed aroud 275mbps and read performance around 375mbps. 

IOPS and latency also as expected for a unit like this; average around 700/800 and latency no higher than 35ms.

I also ensured the latest and correct drivers for the 10Gbps connections on the node; and manually added 8x iSCSI connections; using Windows MPIO (not HP's DSM).

Here are some screenshots.

I hope this helps anyone coming across the thread. 

I will re-post once I have my 4x 400gb SSD drives in the unit and tiering is active. I hope to see IOPS and transfer rates to increase considerably. 

[sebastianmair]

just give the unit back, we did this already at a customer.


kind regards ,

sebastian

This Email has been sent from my IPhone 6, please excuse typos.

[AndreyVV]

Has anyone found a solution to the problem?