Hi, I don't think increasing the mon_osd_down_out_interval timeout alone will really help you in this situation, I remember an older thread about that but couldn't find it. What you could test is setting the nodown flag (ceph osd set nodown) to prevent flapping OSDs, but that's not a real solution. But depending on how often you experience recovery IO the nodown flag could probably make things better until recovery has finished. Zitat von Nicola Mori <nicolamori@aol.com>:

Dear Ceph users,

my cluster is made of very old machines on a Gbit ethernet. I see that sometimes some OSDs are marked down due to slow networking, especially on heavy network load like during recovery. This causes problems, for example PGs keeps being deactivated and activated as the OSDs are marked down and up (at least to my best understanding). So I'd need to know if there is some way to increase the timeout after which an OSD is marked down, to cope with my slow network. Thanks,

Nicola _______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-leave@ceph.io