I have a meta device assigned to a server and that device seems to be saturated during the day time. It shows high i/o, busy % and all that and it seems to be completely saturated. After checking it with EMC, it turned out to be disks on which the members of that meta device (including the meta head) are from four disks which are utilized 100% and EMC has recommended to move whatever part of the application resides on that meta to a different volume.
Now, I have question as to how can I move that part of application to a different volume?
Does that require the application to be stopped?, do I need to ues a BCV and create a copy of that meta and ask the unix team to use the new device instead of the existing meta?
Thanks in advance..
you have a couple of options, for Windows OS you can use "PowerPath migration Enabler" and move to a new device that way. If your OS is HPUX/Solaris/AIX but far the easiest way is to use native logical volume manager and mirror the device.
Thank you for your response.
Its a Unix box and I think it should be mirrored locally by the native tools available.
However, I was thinking in what ways can this be done at storage level (such as use of BCV/clone) so that they have a copy of the original data and they can run their application.
If someone can tell how that cab be done and does it require a downtime at application level?
You don't state what type of Symmetrix or microcode level you are using...
If it's VMAX, then you could simply use the Virtual LUN migration facility to migrate the meta device to a better tier of storage or a different location.
This requires the Symmetrix Optimizer license.
The other alternative is to enable Symmetrix Optimizer and let the array balance out the back end disk utilisation.
Your other alternatives are using host mirror/copy or TF/clone as already discussed.