We can certainly add swap but imo swap is just a slower ram. At least in my
experience when a system starts needing to swap things have already gone
beyond recovery and swap just delays the inevitable program crash. In any
case there's a few ways we can add swap if we really want to:

1) At the job level, create a swap file and activate it for the size needed
for the job.
2) At the jenkins-init level (via jenkins-scripts in releng/builder) we can
add swap for systems here
3) Bake it into the VM via packer file

If we want to try this with a specific job type then I would try it with
option 1 first and see how things go. If we find it really useful we can
move it into 2 or 3.

Regards,
Thanh

On Mon, Nov 13, 2017 at 6:19 PM, Luis Gomez <[email protected]> wrote:

> +1, adding Andy: is it possible to get swap space in the VM from the cloud
> provider? otherwise is it possible to modify the releng/builder packer
> scripts to add the swap?
>
> On Nov 13, 2017, at 3:05 PM, Sam Hague <[email protected]> wrote:
>
> On Thu, Nov 9, 2017 at 5:45 AM, Stephen Kitt <[email protected]> wrote:
>
>> Exactly, that’s what I was alluding to with “the lack of swap probably
>> doesn’t help” ;-). We should really configure VMs with a small amount
>> of swap, it can save the kernel in tricky situations...
>>
> I was looking at this also and wondering about swap. We had a different
> issue where the robot VMs blow up on producing the log.html file because it
> ends up being a 1gb file. With swap I think it would have passed fine. As
> is, we bumped the 2gb vm to a 4gb to make it work. So we will likely hit it
> again.
>
>>
>> On Thu, 9 Nov 2017 01:47:08 -0800
>> Anil Vishnoi <[email protected]> wrote:
>>
>> > I suspect you might hit it again if you run it a bit longer because
>> > of this
>> >
>> > [Thu Nov  2 05:58:08 2017] Free swap  = 0kB
>> > [Thu Nov  2 05:58:08 2017] Total swap = 0kB
>> >
>> >
>> > On Thu, Nov 9, 2017 at 1:39 AM, Stephen Kitt <[email protected]> wrote:
>> >
>> > > On Thu, 9 Nov 2017 10:28:14 +0100
>> > > Robert Varga <[email protected]> wrote:
>> > > > On 02/11/17 23:02, Luis Gomez wrote:
>> > > > > 1) JVM does not kill itself, the OS does instead after the java
>> > > > > process grows to 3.7G in a VM of 4G RAM  (note Xmx is set to 2G
>> > > > > but still the jvm goes far beyond that).
>> > > >
>> > > > Indicates this lies out side of heap -- check thread count.
>> > >
>> > > We verified separately that this is an OOM issue, but one detected
>> > > by the kernel rather than by the JVM (the OOM killer kills the JVM,
>> > > see
>> > > https://jira.opendaylight.org/secure/attachment/14207/dmesg.log.txt
>> > > for details; the number of threads wasn’t an issue here, but the
>> > > lack of swap probably didn’t help).
>> > >
>> > > Upgrading to OpenJDK 8 patch 151 fixed the problem, it might have
>> > > been related to one of the several memory usage bugs in 144 that
>> > > were fixed in 151. It’s probably just moving the goalposts though
>> > > since the problem was new — basically, I suspect we recently
>> > > started using a little too much off-heap memory for some reason,
>> > > and the upgrade to 151 reduces the JVM’s memory usage enough to
>> > > make us fit in our VMs again.
>> > >
>> > > Regards,
>> > >
>> > > Stephen
>>
>
_______________________________________________
controller-dev mailing list
[email protected]
https://lists.opendaylight.org/mailman/listinfo/controller-dev

Reply via email to