ceph health detail actually gives me the number of affected OSDs, e.g.:[WRN] DB_DEVICE_STALLED_READ_ALERT: 13 OSD(s) experiencing stalled read in db device of BlueFS
osd.28 observed stalled read indications in DB device
I tried to inspect the output of ceph tell osd.28 dump_historic_ops but
it's not so easy for me. What I noticed is that no operation seems to
take too much time; the maximum value reported by ceph tell osd.6
dump_historic_ops | grep duration is 1.9306359419999999 (seconds, I
guess). Can this be enough to trigger the warning?
Attachment:
smime.p7s
Description: S/MIME Cryptographic Signature
_______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
