Server Reboot required or not if we extend the LUN size in Linux (HP-MS Service Guard cluster)

[kamalakkannan]

Hi All,
I have a quick question.we have the server running in RHEL 5 OS with HP-MS Service guard cluster ( primary and failover node )with oracle 10g database .we are having the issue that filesystem got full as well as no space in LUN also (HP EVA Storage).So we plan to extend the LUN , then extend the filesystem
Here is the case ,If we extend the LUN size in the Storage level Would it be the server reboot ( on the primary side or in failover side) required in RHEL 5 OS ?

Thanks
Kamal

[Matti_Kurkela]

It will usually be easier to add new LUNs to a cluster volume group than to expand existing LUNs.

Adding new LUNs can be done without a service interruption, even on the primary side. Just add them to your VG, lvextend the LV, then resize2fs the filesystem (in RHEL 5 it automatically detects that this will be an on-line extension).

Then make sure the failover node also detects the new LUNs (disk devices, multipath if you have it) and just run "vgscan" on the failover node to make it detect that the VG has grown.

I'm not sure if the expansion of existing LUNs would be detected correctly without a reboot or unpresenting the disks for a short while, so that the hot-plug subsystem can re-detect their new sizes when they come back. This would require halting the package.

You would then need the "pvresize" command to make the new size known to the LVM, then proceed with the VG+filesystem expansion as usual.

MK