On Wed, Oct 29, 2008 at 05:44:31AM +0100, Francesco Pietra wrote: > Major mistake (shutting down the ups unit) while doing a parallel > computation on all 8 processors (UMA-type machine amd64 lenny with > raid 1 two disks). > > On restarting the machine and the computation, > > "fdisk -l" showed > "Disk /dev/md6 doesn't contain a valid partition table" > > > "df -h" showed > 97G 1.1G 91G 2% /home > (which should have been ca 70% used). > > All other partitions were OK, as they should be.
It would help us to follow if you provided a map of device to mount-point since, e.g. whe don't know what device is /home. > "top -i" showed > all 8 instances fo the parallel procedure > md6_resync (CPU% 6) > md5_resync (CPU% 0) > kjournald > > > After some time md6_resync, md5_resync, kjournald disappeared, leaving > only the 8 parallel procedures. > > Commanding "cat density6.out" the parallel procedure seems to work > regularly, confirmed by "la -l" which shows all files (some as large > as 100MB) for the present computation and the 5 analogous computations > already carried out (of these five computations I had scp sent a copy > to my desktop). > > Where are the data for the current 6th computation being stored? > > The computer was then shutdown and restarted > > df -h output as above. > > fdisk -l > Neither /dev/md6 nor /dev/md5 have a valid partition table. > > cd to my home shows all directories and files for work (in my home > only data from computations, all applications installed in > /usr/local). > > What could I do to set up the sytem in order before resuming the computation? I don't suppose that you kept a copy of your partition table as part of your backup script? e.g. /sbin/sfdisk -d <device>? If so, you can write it back. Doug. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
