| Sorry for the delay, I've been out of the plant.
The file I need to restore is /var/spool/cron/crontabs/root.
Using the nwrecover interface I select the file and appropriate
browsetime. I hit start and the following message comes up
immediately:
Recovering 1 file from /usr/var/spool/cron/crontabs/ into /usr/var/spool/cron
Total estimated disk space needed for recover is 3 KB.
Volumes needed (all on-line):
ilosrv.0049 at /dev/nrmt0h
Requesting 1 file, this may take a while...
Recover failed from NSR server `ilosrv.ilo.dec.com': bad last fragment
Didn't recover requested file ./root
Received 0 files from NSR server `ilosrv.ilo.dec.com'
Recover completion time: Thu Feb 06 11:51:06 1997
Using the nwrecover interface I select the parent directory (crontabs),
use relocate (temp) and appropriate browsetime. Hit start and the following comes up
after a few minutes.
Recovering 17 files within /usr/var/spool/cron/ into /usr/var/spool/cron/temp
Total estimated disk space needed for recover is 32 KB.
Volumes needed (all on-line):
ilosrv.0049 at /dev/nrmt0h
Requesting 17 files, this may take a while...
[uasm: -r ./crontabs/.new..adm]
compressasm -r ./crontabs/.new..adm
[uasm: -r ./crontabs/.new..root]
[uasm: -r ./crontabs/.new..sys]
compressasm -r ./crontabs/.new..root
compressasm -r ./crontabs/.new..sys
[uasm: -r ./crontabs/cronuucp]
[uasm: -r ./crontabs/uucp]
[uasm: -r ./crontabs/.proto..adm]
compressasm -r ./crontabs/cronuucp
compressasm -r ./crontabs/uucp
[uasm: -r ./crontabs/.proto..root]
[uasm: -r ./crontabs/.proto..sys]
compressasm -r ./crontabs/.proto..adm
compressasm -r ./crontabs/.proto..root
compressasm -r ./crontabs/.proto..sys
nwrecover: unknown error running compressasm on file ./crontabs/.proto..sys
nwrecover: recover failed: Error receiving files from NSR server `ilosrv.ilo.dec
.com'
Recover completion time: Thu Feb 06 11:50:11 1997
The files listed above are restored, but with zero byte size.
|
| Couldn't get it to fail... tried
root@mrcofe> echo "compressasm: * ?.*" > /usr/var/spool/cron/.nsr
root@mrcofe> save -smrcofe -bDefault -v /usr/var/spool/cron/
which saved using compression OK.
then
root@mrcofe> recover -smrcofe -v /var/spool/cron
recover: Current working directory is /var/spool/cron/
recover> add *
/var/spool/cron/crontabs
14 file(s) marked for recovery
recover> vol
Volumes needed (all on-line):
mrcofe.01 at /dev/nrmt0h
recover> reloc /tmp
recover> recover
--- but that didn't recover using compression
so tried GUI (nwrecover), and that also wroked but did not indicate compression
and now I've run out of time to experiment....
Out of curiousity, can you try this without compression?? and does it work OK?
|
| I looked up bad last frag in the code tree, and found it in the recover
(mm_recov.c) section (no suprises there), and tried to understand
what was happening to get to this error message.
There is a loop processing "fragments", and if the fragment has too few or too
many bytes, it is tossed, but those that are "just right" are setup with
a volume. When the loop ends a check is made to see if the start frag
is as expected, or the end frag is as expected...if not, an error message
is displayed (bad last fragment).
My take on this means there is bad data coming in from tape, and causing
this. What version of NetWorker did you use to save this data?
Can you see this using NetWorker V4.2a (that is, save and then restore
using NSR 4.2A)?
Thanks for your help on this.
|