HPE StoreVirtual Storage / LeftHand
cancel
Showing results for 
Search instead for 
Did you mean: 

P4000 Storage Managers reporting "No Quorum".

Thomas Mill
Occasional Contributor

P4000 Storage Managers reporting "No Quorum".

Hi all,

I have a 20 node P4000\P4500 array, 2 Managers in 1 site, 2 in remote site and 3rd Failover Manager on 3rd site (Quorum). (SANIQ 8.5 across the board).

The 4 Managers on the actual LHN Storage are reporting "no quorum" over the last 13 hours....(periodically that is, it seems to be reporting in baches, and then hours without any reports).

The quorum server is accessible, and the Failover VM doesn't seem to have restarted recently so I'm trying to find an explanation for this. No updates applied or patches installed in recent months. Any ideas?

I wouldn;pt normally be too concerned but we had a complete system outage on 24 December 2009, so I'm a little twitchy ;-).

While I'm at it, does anyone have a decent document that explains how Lefthand behaves when it loses the quorum Manager, loses a site etc...I still have to work through it when I try to explain how we avoid split brain for instance.

Tom
4 REPLIES
Uwe Zessin
Honored Contributor

Re: P4000 Storage Managers reporting "No Quorum".

Quorum is a value calculated by the number of managers + the FOM (each one has one 'vote').
V Q
1 1
2 2
3 2
4 3
5 3

If the number of votes drops below the quorum value, you should see "no quorum". It sounds like there were/are network disruptions which isolate nodes for a period of time.
.
Thomas Mill
Occasional Contributor

Re: P4000 Storage Managers reporting "No Quorum".

Cheers for responding.

OK, so if the votes drop below quorum then you see a problem. In order for that to happen in our environment (4 Managers and 1 FOM), it suggests that we would have had to have lost visibility of 3 nodes. I thought that losing that many managers would cause an entire outage?

i.e. if we lost access to the FOM (3rd site) and both managers at the remote site (2nd site), would the "available" site (1st Site) not assume it had a problem and therefore also shut down?
Uwe Zessin
Honored Contributor

Re: P4000 Storage Managers reporting "No Quorum".

Quorum loss results in a stall, not in a 'shutdown'. Once quorum is resumed, operation can continue and a resync be started.
.
Terry.giblin
Frequent Advisor

Re: P4000 Storage Managers reporting "No Quorum".

Dear Thomas,

"no quorum" - Dont panic!

Just "Don't tell him Pike"

I might be old fashion, but when ever I see no quorum, I usually panic a little.

As stated previously, it's normally a networking problem.

But I wouldn't wait to find out.

I would be checking everything.

You do not want to lose quorum.

Check your logs immediately and send as much info as possible.

When's the last time you checked your cables.

Terry Giblin