3PAR OS 3.3.1 MU1 is released

[Chris212]

3Par 3.3.1.269-EMU1 available for update on my VSP's. I have not found the release notes... (or it is not available yet)

[ailean]

3Par 3.3.1.269-EMU1 available for update on my VSP's. I have not found the release notes... (or it is not available yet)

Try https://support.hpe.com/hpsc/doc/public/display?sp4ts.oid=1008793289&docLocale=en_US&docId=emr_na-a00027069en_us

[Patrick]

Had three 7000 series arrays scheduled for upgrading to 3.3.1 EMU1 next week to receive a call from HPE this morning that it will no longer be available on the 7000 or 10000 series arrays to any of the 3.3.1 code and the recommended upgrade path is now 3.2.2 MU6.

[Noss]

That's interesting to know. That blows my plans.

Did they give any specific reason?

[Chris212]

Weird...
7000s cannot use compression and all issues with 3.3.1 are mostly related to DECO volumes.

Do not understand this choice.

[Patrick]

That's interesting to know. That blows my plans.

Did they give any specific reason?

Nothing specific other than they encountered problems with the current build. Currently pending a patch from the engineering group before they can move forward.

I'm guessing this is likely DECO related bugs again.

[gtmtc]

Weird...
7000s cannot use compression and all issues with 3.3.1 are mostly related to DECO volumes.

Do not understand this choice.

All documented issues, maybe.

We went to 3.3.1 on our 7400 at the beginning of December (we were promised 64TB virtual volumes back in 2015 when we bought the thing apparently) and even though we don't use dedupe or compression, we've still had a few instances where the Windows cluster attached to the storage just... stops. It's as if the storage disappears, but from the 3PAR console, everything looks fine.

Turns out, the node has a "credit page leak" which causes some particular area of space on the node to drop to zero and it seems to silently fail and rather than failing gracefully and handing over Master Node status to another, nothing happens. Manually rebooting the nodes clears the space for a while (and is all HP have suggested - they've put on a script that monitors the nodes, but it doesn't email when it triggers an alert, so it's not overly useful at the moment), but we're waiting for Patch 21 to actually fix this.

[bajorgensen]

Patrick, do you have TDVV version 1?

We hit a bug after upgrade causing controller reboots, but it hits only in very specific circumstances.
(sorry, under NDA here)
The workaround is to migrate to TDVV3 which you want anyway,
but it is time consuming on a 7000 and require coordination between customer and support.
HPE is doing the right thing holding the upgrade if you are candidate for this issue.

If I were you, I'd reschedule upgrade for early/mid February,
or just go for 3.2.2 until the dust settles on 3.3.1.
It is still going through a rough stability phase, wait for MU2 or MU3.
DECO volumes be damned, I expect 100% stability on a storage system this size.

- Anders

[storsnapper]

So, now 3.3.1 code is not available for 7400 series? That is a royal pain as I was in the process of upgrading all of them so I could use the new system reporter fuctionality built in to the SSMC and decom' the management vm. I've got an open request for upgrade on our production box, there were no plans to implement compression or dedupe.

[bajorgensen]

I would not upgrade until 3.3.1 is default version anyway,
but there are many compelling features.
If you need them, just request an upgrade, the hold probably only affects the specific config I mentioned.

- Anders