Service Guard on XPs, any issues ???

[Henrique Silva_3]

Hello everyone.

We have setup MC/SG on a single FC60 before, and it is been working fine.

We are setting up another MC/SG two cluster node on another pair of servers, and have the ability to go to the new XP SAN, which is more reliable, offers better i/o performance, redundunt, etc,etc.

We have Sybase databases on raw devices and Oracle databases on regular file systems.

I was wondering if there are people on this kind of environemnt already, and if they have had any problems with it. Also, during testing of the package before production, we physically pulled the cables from the FC60 to see if the package would shutdown. I assume we can do the same on the XP port, since most everything is hot-swappable these days.

And finally, I have read somewhere some time ago, that there was an issue with data corruption on oracle database files on the XPS. Anybody know of anything here ?

I assume that the setup on the XP will be the same on the FC60, as far as the volume groups are concerned. We simply mount the proper volume groups on the active server and if that goes does, it simply switches over to the other server.

Anyway, if anybody has anything to add here, I would appreciate it.

Cheers,

Henrique Silva

[melvyn burnard · Accepted solution]

XP Raid Arrays and SAN are fully supported in an MC/ServiceGuard configuration, provided you are using HP-UX 11.x and the latest SG version.
Just plan it before installing an configuring, and ensure you have all the latest patches.
Good Luck

[Steven Sim Kok Leong]

Hi,

I have MC/ServiceGuarded SAP servers (on HP-UX 11.00) as well as MC/ServiceGuarded omniback cell servers (on HP-UX 11.00), both using XP256 for centralised disk storage. I am also using Business Copy (BC) for zero-downtime backup of the databases. between the MC/ServiceGuarded SAP servers and the MC/ServiceGuarded omniback servers.

No major issues so far.

Hope this helps. Regards.

Steven Sim Kok Leong

[Duncan Edmonstone]

Shouldn't be any major issues with this, apply the same rules to volume group/logical volume creation as you would for any integrated cache disk array (i.e. disable bad block relocation).

As for data corruptions... That's totally new to me... who told you this? It wasn't your friendly local EMC salesman by any chance ;o)

HTH

Duncan