On 05/13/2015 02:28 PM, Nathan Scott wrote:
----- Original Message -----
----- Original Message -----
On 05/12/2015 05:43 PM, Nathan Scott wrote:
Changes committed to git://git.pcp.io/nathans/pcp.git master
Ah - this looks like a different problem to PR1106. I can see what it is
likely to be though.
Can you try attached patch, see how it fares? Lightly sanity tested only
at this stage.
OK, that patch fixes the reported segfault.
Also noticed containers.state.running was showing way too many instances.
Should be zero or at most one ("docker ps" shows none running).
The indom is "all containers". The metric value shows whether its running
or not (0/1), so I suspect this is probably operating as designed - or are
you saying they have value "1" when they shouldn't?
no, you are right - I was confused about what "all containers" means - it
includes containers that have exited but have not had their image/overlay
removed, which is the default when a container exits. docker run --rm
will remove it when the container exits, which seems backwards to me!
From our chat, sounds like its fine as-is and "docker ps -a" output indeed
matches what we have. Looking forward to hearing more about the profiling
run for pmdaroot with 1000s of containers too - thanks!
still getting the pmdaroot timeout - tweaking this in pmcd.options works
around it, so we can capture a profile.
Lemme know if time is short & I'll prep something for qa from your earlier
recipe (tomorrow).
time is short (have to head out for an appointment). So I'll pick this up
again tonight. In the mean-time might as well commit that patch to your
tree and I'll merge to master tonight. (and craft a qa test from the
recipe)
Cheers
|