If I understand, it's the disk array for the vtapes that failed?
Did you run 'amrmtape CONFIG LABEL' for each label so that amanda know
that the backup on those tape are no longer valid?
If not, then why amanda would do full dump? amanda think the previous
dump are still available.
jsummers wrote:
Hello All,
Man it has been a while since I have had a need to post a question here!
Amanda has been running beautifully and trouble free! That is until my disk
array broke. Fortunately I had another one that I brought online with a new
install of Amanda also.
I am using VTAPES with the hard disk style backups. Got everything running and
I wanted to get caught up on the backups. So I set the runtapes to 8. But for
several days now all that is used are two tapes and I get a message:
partial taper: No space left on device
Each day it is a different DLE.
Looking through the log file I can see where it runs out of tape and switches
to the next one. Which I think explains the partial taper message. Then it
will right a second tape. But that is where it stops. I guess I was thinking
that the planner would configure a run to use all eight tapes if needed.
I have the following set in amanda.conf:
runtapes 8
dumpcycle 2 weeks
runspercycle 14
tapecycle 37
length 40Gb
I did notice that the maxdumpsize was set to -1.
I commented that out so that it will default to ( runtapes * length ).
Do ya'll think that may get it to use the eight tapes if needed?
Nothing is left in the holding disk, so I do not suspect autoflush or anything.
In the report, planner states that many DLE's last full will be overwritten in
2 - X runs. So I think there are some level 0's she could be doing??
Amanda do level 0 in advance only to help it balance the schedule.
You can also force full backup with 'amadmin CONF force ...' command,
but amrmtape is the best solution for you.
Jean-Louis
