On Wed, Aug 12, 2020 at 12:28 PM Sergio Belkin <sebelk@xxxxxxxxx> wrote: > > Hi! > I 've just had a problem using EarlyOOM + ZRAM. I haven't a disk-based swap partition. > I was using mainly Zoom (desktop app) + Firefox + VirtualBox (Debian with 4GB of RAM), and EarlyOOM killed Zoom in the middle of a call :( Hi, Could you open a bug report against earlyoom, and attach 'journalctl -b -o short-monotonic > bugID-journal.txt' It's not necessary to trim it. But if you do, keep about 5-10 minutes prior to the first SIGTERM. I think what's happened is earlyoom issues multiple SIGTERM, which are ignored or delayed in actually terminating by the processes with the most badness. And it just so happens that the zoom process badness reacted to the terminate request despite having an order magnitude lower badness. Off hand I'd say earlyoom probably should only SIGTERM the top 3 offenders, possibly with some delay in between each one. And then once we're below the kill watermark, issue SIGKILL to the top offender again. The difficulty is that's a VM, in this case. We might need an oom_score_adj for qemu and virtualbox, because I think it's a bad idea to kill off running VM's. -- Chris Murphy _______________________________________________ devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxx Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@xxxxxxxxxxxxxxxxxxxxxxx
