Open Stack

On Feb 4, 2013, at 7:58 PM, Clark Boylan <clark.boylan at gmail.com> wrote:

> You may have noticed that earlier today Jenkins jobs were failing
> because logs.openstack.org had run out of disk space and individual
> jobs were unable to copy their logs to the log server. After looking
> at the problem we noticed that after the move to running the entire
> tempest suite in the gate-tempest-devstack-vm jobs had increased total
> log size for each of those jobs to close to 100MB. To bring that
> number back under control the gate-tempest-devstack-vm jobs are no
> longer capturing the OpenStack service screen logs (that data is
> captured by syslog so it was redundant) and devstack-gate-vm-wrap.sh
> is compressing all logs before shipping them to the log server.

The problem with this is syslog compresses tracebacks onto a single line
and truncates them. If we are depending on syslog only can we up the max
message size so we can still see the full traceback?

$MaxMessageSize 64K (rsyslog)
log_msg_size 65536 (syslog-ng)

Vish