vSphere 6 issues with non Enterprise Grade SSD as Storage?

[mervincm]

I am wondering if anyone else is seeing any strangeness when using non enterprise grade SSD storage with the freshly released vSphere 6?

Specifically, I am having problems adding them as local storage. I have not had any issues with my enterprise Hitachi HDD, nor my elderly Intel SLC SSD (X25E) They show up appropriately as HDD and SSD datastores, and when I add them using add storage, it happens quickly and w/o incident. Also, my consumer grade Seagate Baracuda green HDDs were added w/o incident.

When I tried to add a consumer SSD (240GB OCZ Agility 4) it took a few minutes to add it (instead of the expected few moments) and then the interface was "refreshing" for a few more minutes, then it simply disappeared.

This system (Lenovo TS-140) was working as built just fine with 5.5, but after a fresh install (not upgrade) to v6 the OCZ SSD is not functional.

I have a second system (another TS140) that is having similar issues. The Intel SLC SSD is added quickly and w/o incident. The Hitachi HDD, as well, no issues. I added an unused OCZ 60 GB agility 3 consumer grade SSD, and while it took a long while to add it in, it seems to work, OK

However, the two corsair consumer SSD (120GB force) well if you try to add either, it errors out, does not finish, and you need to hard power off the host to get it to respond again.

This second system did NOT have any SSDs in it when it was running 5.5, but they do pass the corsair toolkit tests.

 

[mervincm]

On the first system, here are some relevant logs

File system [vSphere2-240-OCZ,
5522d671-7f225d2c-bc10-fc4dd4f579ec] on
volume 5522d54d-952dd810-29da-fc4dd4f579ec
has been mounted in rw mode on this host.
info
2015-04-06 12:54:41 PM

Lost access to volume
5522d671-7f225d2c-bc10-fc4dd4f579ec
(vSphere2-240-OCZ) due to connectivity issues.
Recovery attempt is in progress and outcome will
be reported shortly.
info
2015-04-06 12:58:05 PM

Device t10.ATA_____OCZ2DAGILITY4_________
___________________OCZ2DOAO8C25M0CKS-
RQF3 has been removed or is permanently
inaccessible. Affected datastores (if any):
Unknown.
error
2015-04-06 12:58:05 PM


Permanently inaccessible device
t10.ATA_____OCZ2DAGILITY4_______________
_____________OCZ2DOAO8C25M0CKSRQF3 has
no more opens. It is now safe to unmount
datastores (if any) Unknown and delete the
device.
info
2015-04-06 12:58:05 PM

 

[Biren78]

No good! The worst drives I've tried with 6 are the sandisk cloudspeed eco and intel s3500. They worked fine.

 

[CreoleLakerFan]

Device t10.ATA_____OCZ2DAGILITY4_________
___________________OCZ2DOAO8C25M0CKS-
RQF3

There's your problem - you're using OCZ Agility drives. Those are the friggin worst ... My first SSD was an Agility ... I had to RMA the damn thing four times - they finally replaced it with a Vertex, which is only slightly less shitty.

OCZ is the worst OEM. Ever. As evidenced by the fact that they no longer exist as a company.

 

[namike]

Just to share my experience, I did an in-place upgrade of my home ESXi box from 5.5 to 6.0 with a Crucial M4 128GB as my data store with no issue.

 

[Deci]

I had an issue with formatting an ssd as a data store, the trick was to just leave it be for about 10 minutes and then it came up fine, Hdd light was solid for the entire time pretty much and the host interface was responding very slowly, after it finished it has been working flawlessly since.

 

[mervincm]

There's your problem - you're using OCZ Agility drives. Those are the friggin worst ... My first SSD was an Agility ... I had to RMA the damn thing four times - they finally replaced it with a Vertex, which is only slightly less shitty.

OCZ is the worst OEM. Ever. As evidenced by the fact that they no longer exist as a company.

Well of the two OCZ drives, the agility3-60 is working fine (after a SLOW add) and the Agility4-240 failed out after an attempted add. Compare that to the two corsair force-120 drives that I had to REBOOT the host to recover.

OCZ seems to be working better than corsair here

and the OCZ Agility 4-240 was working just fine under 5.5.

 

[mervincm]

I had an issue with formatting an ssd as a data store, the trick was to just leave it be for about 10 minutes and then it came up fine, Hdd light was solid for the entire time pretty much and the host interface was responding very slowly, after it finished it has been working flawlessly since.

Yes I had a similar experience. with the corsair force 120's the HDD drive was solid and I suppose I didn't wait long enough, but I am pretty sure I waited at least ten minutes. With the Agility4-240, it worked after about 5, then the storage interface was busy for 5 then it removed the mounted drive.

Right now I am running long smart tests on them all via my test box.

 

[TeeJayHoward]

Just for another data point, Samsung 840 Pro drives here are all working great. (They're still technically consumer drives, right? They're not SAS.) I'm using my 850 Pros for flash cache, otherwise I'd report on those too.

 

[Patriot]

Just for another data point, Samsung 840 Pro drives here are all working great. (They're still technically consumer drives, right? They're not SAS.) I'm using my 850 Pros for flash cache, otherwise I'd report on those too.

Yes, those are prosumer drives. Samsung makes DC versions of them with PLP for entry level enterprise drives.

 

[mervincm]

Extended SMART test completed on both corsair 120 GB w/o issue. OCZ 240 didn't seem to want to run a smart test under parted magic, so I tried it on my windows machine, seems to be working there so far as well....

I might pick up an 850 pro to test, but for now I am just running everything off of the NFS share on my Synology.

I appreciate the other data points. I reviewed my notes and I forgot to mention that I also did a server Firmware upgrade just before the move to 6.0. I suppose that is a possible problem source as well.