Hi Jan,

Thanks for the patch. It definitely gets me a lot further along, but my
backups are confounded by a perhaps
even stranger problem --

I run the backup.sh script and see this in the backup log:

Partition /vicepa: inodes in use: 133, total: 2097152.
Partition /vicepb: inodes in use: 997, total: 2097152.

Date: Thu 08/24/2006

22:43:18 Partition /vicepa: 63091568K available (minfree=4%), 46960816K
free.
22:43:18 Partition /vicepb: 131018296K available (minfree=4%), 126483880K
free.
22:43:18 Partition /backup: 30901448K available (minfree=4%), 30901440K
free.
22:43:18 VLDBLookup: VLDB_size unset. Calling VCheckVLDB()
22:43:18 VLDB_Lookup: cannot find "01000001"
22:43:18 Volume replica 1000001 doesn't exist!

22:43:18 Skipping backup for volume 7f000000

22:43:18 VLDB_Lookup: cannot find "01000002"
22:43:18 Volume replica 1000002 doesn't exist!

22:43:18 Skipping backup for volume 7f000001

... (every volume in /vice/db/dumplist)

22:43:18 Skipping backup for volume 7f000011

22:43:18 VLDB_Lookup: cannot find "0100000d"
22:43:18 Volume replica 100000d doesn't exist!

22:43:18 Skipping backup for volume 7f000012

22:43:18 VLDB_Lookup: cannot find "0100000e"
22:43:18 Volume replica 100000e doesn't exist!

22:43:18 Skipping backup for volume 7f000013

22:43:18
22:43:18 Attempting to retry any failed operations.
22:43:18
22:43:18 Successfully backed-up Volumes:
22:43:18
22:43:18 Only partially successfully backed-up Volumes:
22:43:18
22:43:18 Volumes that FAILED backup:
22:43:18
22:43:18 Volumes that were NOT backed-up:
22:43:18 0x7f000000                     /
22:43:18 0x7f000001                     diablonet.user
22:43:18 0x7f000003                     diablonet.group
22:43:18 0x7f000004                     diablonet.sys

... (every volume in /vice/db/dumplist)

22:43:18 0x7f000012                     user.rdekema
22:43:18 0x7f000013                     group.searches
22:43:18 Histogram of sizes of dump files

Good samples:     count = 0    mean = 0   stddev = 0   90%CI = 0

Bad samples:      underflow = 0    overflow = 0

22:43:18 Histogram of DataRates for transfer of dump files

Good samples:     count = 0    mean = 0   stddev = 0   90%CI = 0

Bad samples:      underflow = 0    overflow = 0

22:43:18

Running tape.pl with  backupdir /backup, dumpdir /backup/24Aug2006,
tape /dev/nst0 on host blossom.diablonet.net.
Label: Coda I Thursday


and that's it. But yet, I created backup volumes, and they show up the VLDB:

blossom.diablonet.net> strings VLDB | grep backup
src.coda.0.backup
src.gnu.0.backup
user.rdekema.0.backup

... and so on.

I produced the backup volumes using the procedure,

volutil -h <master-replica-server-name> lock <volid>
volutil -h <master-replica-server-name> backup <volid>

for each volume, then when I did this for all volumes, ran bldvldb.sh. Is
this
anywhere close to what I am supposed to be doing to set up backup volumes?

What would be a good strategy to debug this? Is there any sort of debuglevel
flag for the backup program? I don't see one in the manpage but one never
knows. :)

Thanks, Sean
scaron_at_umich.edu