[linux-lvm] What happens when a disk dies?

[Adrian Phillips]

>>>>> "Kyle" == Kyle S Hoyt <Kyle_S_Hoyt at raytheon.com> writes:

    Kyle> I am very interested in starting to use LVM.  We have large
    Kyle> amounts of data that is fairly dynamic.  Since it is
    Kyle> dynamic, we don't need to incur the cost expensive SCSI RAID
    Kyle> solutions for data integrity.  We currently use four 80 Gig
    Kyle> Maxtor Hard Drives to store our data at test events. Of
    Kyle> course this means with a /drive1 /drive2 /drive3 and a
    Kyle> /drive4 directory.  So the user has to search the drives to
    Kyle> find the file of interest.  We kind of hide this by creating
    Kyle> links but the links need to be periodically updated as files
    Kyle> are added or deleted.  LVM solves this problem.  But before
    Kyle> I move to LVM, I curious on what happens if a drive dies.

    Kyle>  If drive2 goes down and I replace it with another drive, do
    Kyle> I just lose the data on drive2 or do I lose all of the data
    Kyle> in the Logical Volume?

This is a filesystem thing not LVM (well LVM is involved if you use
it). If you lose a drive the filesystem is going to be unhappy to put
it mildly.

I don't understand your problem here , you can :-

- mount each disk under /this_is_the_data as its own filesystem (not
  using LVM necessarily), say /this_is_the_data/drive<n>

- search from /this_is_the_data

- if a disk goes then umount drive<n> and searches can still continue

If this does not work then more information is needed.

Sincerely,

Adrian Phillips

-- 
Your mouse has moved.
Windows NT must be restarted for the change to take effect.
Reboot now?  [OK]