Is it possible to resize a filesystem by resizing the LUN on RHEL 6.4 64-bit with LVM and no impact to running applications? The research I have done so far seems to take the approach of adding a new LUN and then expaning the volume group to the new LUN. I'm looking for an approach that avoids a new LUN this as we virtualize all of our Linux. Not requiring a reboot is very important.
you can change the filesystem while it is running but you are asking for a disaster to happen ... essentially, you will be creating files and then messing up the filesystem underneath it ... it would be similar to jumping into a sand pit expecting a fairly solid spot but getting quick sand instead ...
i strongly advise against resizing filesystems while they are running or in use but if you really want to try it -- try testing the process you envision on a "disposable" server with LVM enabled and confirm if the risks involved are acceptable ...
There is alot of work so be sure to read the entire thing.
Test it on some test environment multiple times.
I would not recommend it tho, since adding a new lun is much safer/easier operation and in these days with tiering and stuff going on on the storage performance wise probably the same.
It would be simpler to add a LUN to the OS and then add that to the volume group. You can then work with the filesystems cleanly, even if they are mounted. Is there a reason you cannot allocate another LUN to the virtual server? I'm told by our VMWare man that we have a limited number of LUNs he can attach.
If you are trying to keep it simple for neatness, could you add a much larger LUN, add it to the volume group, use pvmove to migrate the data from the existing LUN to the new one and release the old LUN? That way you will end up with the same number of LUNs attached.
there never was a question about being able to resize a filesystem while it is running ... given that a reboot is to be avoided as much as possible, i am assuming that the processes in question are hyper-critical ...
so the issue is more -- what are the consequences of a botched process and can the company afford it? or put in another way, can you or your manager afford the risks to production?