-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 Unfortunately some of the hosts affected most by this problem are LXC servers. They rely upon "cgroup_enable=memory swapaccount=1" on the kernel command line, making sure that I can start containers with memory restrictions. I have to make sure that these greedy Java apps don't put other containers at a disadvantage.
I don't care about systemd here. Theses hosts run with good ol' sysvinit. But its worth mentioning that I saw the "cpu got stuck for 22s" on Linux desktops as well, not related to LXC at all. They use systemd, though, i.e. cgroups are still in the loop. There was no such problem with 4.7.x. Do you think that cgroups decrease the stability of 4.8.x and 4.9.x in general? Regards Harri -----BEGIN PGP SIGNATURE----- iQEzBAEBCAAdFiEEH2V614LbR/u1O+a1Cp4qnmbTgcsFAliPtX4ACgkQCp4qnmbT gcvN/wf+Ism+hQyLR26YKw4NInzfxGkr2CpiED5q4iWWPaijUtVymp85VBsWIsFa Ud4w3qHca1XCryjqiTAdK6a14DCzq6Vtp7Go5vI6tBVjWoAQlKKbDdf0iesxSPbE s/TIAeOxCr9qPOiGi82tIUCBrPSZW14bzHdZER0dHPUJkEM/Kmi9qY31Z00CY1FW vMPcM6/E3medr/MUyTVB9+v7I8N6ReKICYiDtNNm87WqpcY7X3QX/Lfz3mNlXK3D 8M2cvN73hPNone9Int/XGWE696v4Fh3coSYzo0m4wdW7MJxri2/aFki4i3T6uJkw K7873UtpGy8FmqW7c5tcZmy9Mt/Hag== =Cs62 -----END PGP SIGNATURE-----
