Web lists-archives.com

Amanda since the upgrade to Stretch


Since the upgrade from Jessie to Stretch, I have noticed occasional 
misbehaviour in my amanda setup. I didn't change anything in my hardware 
or software configuration that might have caused it. (Famous Last 

The most recent occurrence was last night, and it was typical of what 
happens when it goes wrong:


These dumps were to tape RealBackup08.
Not using all tapes because taperflush criteria not met.
There are 210G of dumps left in the holding disk.
Run amflush to flush them to tape.

The next 2 tapes Amanda expects to use are: RealBackup09, RealBackup10.

  localhost /opt lev 0  partial taper: taperflush criteria not met


Running amflush RealBackup as the backup user works fine to flush the 
dump. My question is why didn't the flush work during the main overnight 
backup run?

The system is otherwise idle during the backup run. I have it set to run 
at 1:30am and I was in bed last night by about 00:20. I had shut down 
both VirtualBox VMs on my system, and the backup script shuts down a 
bunch of services like mysql (actually mariadb nowadays) and svnserve on 
the machine before commencing the backup, so as to get as good quality 
as possible a backup. So the machine shouldn't have been doing anything 
else at the time except servicing the backup.

My amanda "tape" is a set of 10 vtapes on a 3TB hard disk mounted in an 
external USB drive cage. The holding disk is the second 3TB disk in the 
cage. Both disks are mounted by me after booting and left mounted until 
the next reboot. This issue I am seeing happens only a small percentage 
of backup runs. I backup every weeknight. This configuration worked 
without ever reporting problems in Jessie.

What does "taperflush criteria not met" mean?