cron executing jobs twice, save-entropy errors

I found a strange error with my jails (ezjails) producing double status emails.
The apparent error is similar to the one mentioned here.

As I investigated further I found out that "sometimes" the ezjail runs the cronjob twice - so creating a race condition with two save-entropy jobs trying to move and alter the files concurrently. Or sending out a status email twice...

/etc/crontab looks normal, so no double entries or so.
/var/log/cron shows the duplicate starts of the jobs.

What can cause such a behaviour? As cron is triggered via timestamps, could it be possible that under high load (define high here) time is "too slow"?
 
ezjail-admin update missing

It seems I solved the problem.
I missed an
# ezjail-admin update -u
to update the base-jail (binaries linked into all other jails).
So there was a difference between the system binaries of the host-system and the basejail.

So, after doing a
# freebsd-update fetch && freebsd-update install
don't forget to do an
# ezjail-admin update -u
to update your ezjails to the same level as the base-system.

(ezjail-admin update -u: -u uses the freebsd-update mechanism)


Will wait another day to see if it is solved.
 
Problem solved by issuing the commands mentioned above.
Errors occurred (IMHO) due to mismatch of system binaries in host- and (ezjail-)basejail-system.
 
Back
Top