EMERGENCY: killing pvmove
Hello List,
I'm in a emergency situation so I haven't had the time to go through all the threads. I have had a strange problem and through my interaction it is now more severe as I think, so I have to get advice quickly.
I have several RAID subsets as physical volumes, each 1,4 TB in size, combine to a volume group of 8,6 TB. This VG is splitted in 6 LV's. One of the LV I did resize last week, and therefore added one PV of again 1,4TB to the VG, and resized one LV.
After that -- everything went ok, even resizing the ext3 partition -- I got Input/Output errors while writing to the newly resized partition (writing to the extended part of the partition). As realising this, I put all the data off this partition, deleted it and was then trying to remove the Raid subsystem with pvmove. While doing "pvmove /dev/sdg" I just got an "segmentation fault".
As this wouldn't be enough, I restarted the computer putting the new RAID subsystem on a different scsi controller, as I thought this could be the problem. Then the pvmove command worked!
But as I realised shortly, I just moved the wrong one, as /dev/sdg is now another subsystem (the damaged one is now /dev/sde). Now I have the Problem that my sane part of the data is slowly moving to the insane disk, which I can't stop! Even ctrl-c does not help...
How to stop it? Is a SIGTERM or SIGKILL allowed? Pleas help, I'm running out of time!
Thanks a lot!!!!
yours
Sebastian
pvmove --abort
Thanks for your fast answer!
Quoted 2 lines · Alasdair G Kergon
pvmove --abort
[root@mex root]# pvmove --abort
pvmove: unrecognized option `--abort'
pvmove [-A|--autobackup {y|n}] [-d/--debug]
[-f|--force] [-i|--ignore_read_errors] [-h|--help] [-t|--test]
[-v[v]|--verbose [--verbose]]
[--version]
[{-n|--name} LogicalVolume[:LogicalExtent[-LogicalExtent]...]] SourcePhysicalVolume[:PhysicalExtent[-PhysicalExtent]...]} [DestinationPhysicalVolume[:PhysicalExtent[-PhysicalExtent]...]...]
[root@mex root]#
Is it because I'm using lvm1 ???