On 19/12/2018 23:15, Ivan Babrou wrote: > We're sticking with the following patch that allows runtime switching > between XFS memory reclaim strategies: > > * https://github.com/bobrik/linux/pull/2 > > There are some tests and graphs describing the issue and how it can be solved. > > Let me know if you think this can be incorporated upstream, I'm fine if not. This hack and fs.xfs.memory_reclaim=0 solved my problem where backup server (doing rsnapshot, so tons of rsync+hardlinks+rm+cp of milions files) was barely usable. Things like ssh or other network services were stuck for 10-30 minutes doing reclaim. Hack but for now server behaves much better than with vanilla kernel on this type of workload. Thanks, (resend) -- Arkadiusz Miśkiewicz, arekm / ( maven.pl | pld-linux.org )
