watanabe@nttcvg.NTT.JP (Yasuhiko Watanabe) (06/15/91)
Now, the disk in my system is not stable, so I would try the Backup it by bru. But while "bru"ing, when the location is encountered to (unrecovered) block, the "bru" stops, but I could not kill the "bru" process. ((unrecovered) is printed in /usr/adm/SYSLOG file with meadia location address)Moreover, the system could not down with the disk "umount". So in the next boot, the "fsck" would start (it takes long time for big media). HOW TO STOP "BRU" WHICH STOPED BY THE MEDIA TROUBLE (unrecovered)? hohiji w. -- ------------------------------------------------------------------- Yasuhiko Watanabe, Senior Research Engineer phone: +81-468-59-4735 Visual Perception Laboratory fax: +81-468-59-2898 NTT Human Interface Laboratories 1-2356 Take Yokosuka-Shi Kanagawa 238-03 Japan CSNET: watanabe%nttcvg.NTT.jp@relay.cs.net -------------------------------------------------------------------
olson@anchor.esd.sgi.com (Dave Olson) (06/18/91)
In <1578@nttcvg.ntt.jp> watanabe@nttcvg.NTT.JP (Yasuhiko Watanabe) writes: | Now, the disk in my system is not stable, so I would try the Backup it by | bru. But while "bru"ing, when the location is encountered to (unrecovered) | block, the "bru" stops, but I could not kill the "bru" process. | ((unrecovered) is printed in /usr/adm/SYSLOG file with meadia location address)Moreover, the system could not down with the disk "umount". So in the next | boot, the "fsck" would start (it takes long time for big media). | HOW TO STOP "BRU" WHICH STOPED BY THE MEDIA TROUBLE (unrecovered)? | hohiji w. It is hard to tell for sure, but what seems to have happened is that bru is stuck at an unkillable priority, due to the disk errors. Normally, this shouldn't happen :); bru should get an error from the kernel on its read from disk (or write to tape). I haven't heard of this particular problem before. At a minimum, we would need to know the disk types (scsi, esdi, ipi) and tape type that are in use, as well as the machine type and the OS release to try and figure out what is going on. You should go through the TAC folks if you want to pursue this. However, the short answer is that when you have unrecoverable disk errors, you can wind up with lots of trouble. If there are only a few bad blocks, and you can't back up the disk anyway, you might try using fx to foward the blocks that are bad. Read the COMPLETE fx man page carefully, and if you still don't understand, get an FE or someone who does to do it for you, or you may lose all the data on the disk. As an extreme desperation measure, if your backups are good, and all else fails, you could try re-formatting the disk...