FOM fails to remain online

[Peter J West]

Hi,

I've just finished configuring a multi-site P4000 installation which consists of a total of 6 nodes.

The best practice recommendation is that I create two managers on each site and then have a fifth running as a FOM in a third site.

I've tried installing the FOM under ESX using local storage but it's not at all stable.

The VM boots up and initially the manager starts to run, but then within a few moments it goes offline and you're unable to get it working again (it appears you can't right click a FOM and choose Start Manager like you can on a normal storage node).

This problem also means I can't apply the upgrades which are waiting.  On the FOM console there are some errors being shown relating to dbd_manager?  Is this expected or part of the issue?

I've tried searching for a later copy of the FOM but so far as I can see only v9.0 is available.

Grateful for any help that yoiu can offer.

Regards

Pete

[Jitun]

dbd_manager is the manager service on the unit.
Could you please share the screenshot.

Is the FOM seen in CMC ?
Have you been able to add it to the management group ?
If yes, does it say Manager Starting/Stopping ?

[Bart_Heungens]

First of all: can you ping it? Is the network OK? VLAN-ing?

[Peter J West]

Hi,

The network side of things appears to be ok - I can ping the FOM and there's no dropping of traffic.

This morning when I came in it wouldn't even let me view the configuration of the manager and it simply kept saying it couldn't find a system with that ID (and then displays the MAC address).  I did a restart on the FOM and searched for the system and it found it. 

I've attached an image file to this post that show's the status - initially it comes up fine and it IS part of the management group.  But within 60 seconds it goes offline and then there doesn't appear to be any way to bring it back online.

The second image attached shows the messages on the console.

Any clues would be really welcome. :)

Pete

[Bart_Heungens]

Something strange seems to happen. Never saw this issue before...

Can you remove the FOM from the MG, delete it on ESX level and start all over with a new clean FOM? Re-aasign an IP etc...

If this doesn't help, I would suggest to log a call at HP... Unless other peaople have other ideas...

Kr,

Bart

[Jitun]

Obsolete socket error message can be ignored.

I would suggest the same as Bart.

Remove the FOM from management group.
Delete the FOM from ESX and use a new FOM.

[Peter J West]

Hi again,

Thanks for the suggestions - i'm trying out a couple of things at the moment.

One question though.

If all of my storage nodes are using jumbo frames and flow control then am I correct in assuming that the FOM should also have the same settings?

Thanks

Pete

[Bart_Heungens]

Hi,

If you use Jumbo Frames (which I don't recommend), yes all devices in your SAN network (which is actually a LAN) must be configured for Jumbo Frames...

Kr,

Bart

[Peter J West]

Hi,

Thanks for your comments - but can I ask why you don't recommend jumbo frames?

In our environment we are running the SAN nodes on a totally seperate infrastructure to the data network - of course there is a link - there has to be so we can manage to nodes from our normal data network.  But the VLAN configuration means we don't have SAN traffic crossing over to the normal data network and in this way we can keep the two systems seperate.

I think I may have identified the cause of my problem - i'm testing at the moment and if so then i'll report back with what eventually fixed the issue.

Kind regards

Pete

[Jitun]

If you use Jumbo Frames... every other device in you SAN should support Jumbo Frames.

If you have Jumbo frames enabled on the nodes .. try disabling Jumbo Frames on the switch port...

Manager of that node would go offline... It will not be able to communicate with other Managers.

I recommend using Default Packet as well. The difference is performance is not worth the trouble, unless you have to have Jumbo Frames enabled.