On 03/04/2015 05:07 PM, Simon Déziel wrote:
> Before loosing anyone's time on this issue, I'll first make sure the
> hypervisor's RAM is sane. So I'll get back after some memtest86+ then
> I'll look at the 4.0 kernel.
After 15+ hours of memtest86+ (8 full passes), I don't think the RAM is
at fault. I ran a fsck on every VMs and many of them had issues. I'm
still using the default Ubuntu kernel for now.
One weird thing I noticed is that apparently all the affected VMs have
trouble with the /etc/cron.daily/logrotate job:
On 03/04/2015 05:07 PM, Simon Déziel wrote:
> Before loosing anyone's time on this issue, I'll first make sure the
> hypervisor's RAM is sane. So I'll get back after some memtest86+ then
> I'll look at the 4.0 kernel.
After 15+ hours of memtest86+ (8 full passes), I don't think the RAM is
at fault. I ran a fsck on every VMs and many of them had issues. I'm
still using the default Ubuntu kernel for now.
One weird thing I noticed is that apparently all the affected VMs have daily/logrotate job:
trouble with the /etc/cron.
error: error creating output file /var/log/ syslog. 1.gz: File exists daily/logrotate exited with return code 1
run-parts: /etc/cron.
On those, there was a very old /var/log/syslog.1 and a more recent .1.gz
one too. I removed both and will see if it helps.