I could try cross compiling and running it on Rpi 3 with 1gb RAM just for fun.
I try to keep my ram as low as possible on virtual machines. Maybe a qemu VM behaves differently. I will try that, too. Thanks for the help. Am 13. Sep. 2018, 16:10, um 16:10, Mathieu Lonjaret <[email protected]> schrieb: >Last time I checked (but it was a long time ago), RPis were very low >on memory, and doing anything on them was very slow. For example, I >think even building Perkeep itself was not possible. So I wouldn't get >my hopes up for Perkeep on an RPi. > >On Thu, 13 Sep 2018 at 12:08, Niklas Merz <[email protected]> wrote: >> >> Witj 4GB of RAM delegated to this container it runs more stable. >Seems a bit much to me, but maybe this is only the case with LXC >containers in Proxmox. Maybe I will try running perkeep on metal or a >Raspberry Pi to check this. >> >> Am Sonntag, 2. September 2018 02:34:07 UTC+2 schrieb euankemp: >>> >>> > If you are constraining your container so it has a tiny amount of >memory, that would certainly explain why the kernel is OOM killing it. >>> >>> Note that we can already answer that from the dmesg output: >>> >>> [9501543.900845] Memory cgroup out of memory: Kill process 22465 >(perkeepd) score 530 or sacrifice child >>> [9501543.901793] Killed process 22465 (perkeepd) total-vm:1648956kB, >anon-rss:1062200kB, file-rss:0kB, shmem-rss:0kB >>> >>> The fact that it says "Memory cgroup out of memory" not "Out of >memory" at the beginning indeed means it's a cgroup limit. We can also >see that it was using about 1GB of residual memory, so it's reasonable >to assume the cgroup has a memory limit of greater than 1GB. >>> >>> That's not enough to answer whether this is a memory leak or not >though. 1GiB seems like a lot of memory for perkeep, but it also might >be okay for some workloads, so without information about what that >memory's being used for, we can't be sure whether anything's really >wrong or not. >>> >>> On Sat, Sep 1, 2018 at 5:27 PM Theodore Y. Ts'o <[email protected]> >wrote: >>>> >>>> On Sat, Sep 01, 2018 at 11:20:21AM -0700, Niklas Merz wrote: >>>> > Can I provide you something to test this? I am using a Debian 9 >in a LXC >>>> > container on Proxmox with a very simple systemd script. >>>> >>>> Are you setting up an explicit or implicit memory limit on your >>>> containers? >>>> >>>> If you are constraining your container so it has a tiny amount of >>>> memory, that would certainly explain why the kernel is OOM killing >it. >>>> >>>> Quoting from: >https://stgraber.org/2016/03/26/lxd-2-0-resource-control-412/ >>>> >>>> All limits can also be inherited through profiles in which case >>>> each affected container will be constrained by that limit. That >>>> is, if you set limits.memory=256MB in the default profile, >every >>>> container using the default profile (typically all of them) >will >>>> have a memory limit of 256MB. >>>> >>>> - Ted >>>> >>>> -- >>>> You received this message because you are subscribed to the Google >Groups "Perkeep" group. >>>> To unsubscribe from this group and stop receiving emails from it, >send an email to [email protected]. >>>> For more options, visit https://groups.google.com/d/optout. >> >> -- >> You received this message because you are subscribed to the Google >Groups "Perkeep" group. >> To unsubscribe from this group and stop receiving emails from it, >send an email to [email protected]. >> For more options, visit https://groups.google.com/d/optout. > >-- >You received this message because you are subscribed to a topic in the >Google Groups "Perkeep" group. >To unsubscribe from this topic, visit >https://groups.google.com/d/topic/perkeep/DUmc92gUocI/unsubscribe. >To unsubscribe from this group and all its topics, send an email to >[email protected]. >For more options, visit https://groups.google.com/d/optout. -- You received this message because you are subscribed to the Google Groups "Perkeep" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
