|
Hi,
I've been trying the new multipath configuration options in RHEL8 beta, the options are listed in https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/8-beta/html-single/8.0_beta_release_notes/index#file_systems_and_storage_2.
Below is my multipath configuration:
defaults {
user_friendly_names yes
find_multipaths yes
marginal_path_double_failed_time 120
marginal_path_err_sample_time 121
marginal_path_err_rate_threshold 10
marginal_path_err_recheck_gap_time 30
}
and this is the multipath storage:
3600124f0f822cba90003551f67355d000d dm-2 SUN,ZFS Storage 7120
size=100G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
`-+- policy='service-time 0' prio=50 status=active
|- 6:0:0:4 sdf 8:80 failed faulty running
|- 6:0:1:4 sdbv 68:144 failed faulty running
|- 8:0:0:4 sdab 65:176 active ready running
`- 8:0:1:4 sdcy 70:96 active ready running
To do the test, I use dd to write to the storage and use "iostat 1" to check IO, then log onto the fc switch to disable a port, after port being disabled, two paths will fail, and IO is only seen on the remaining two paths, then I immediately re-enable the
port, and IO will again be seen on all 4 paths, and then I do this path disable and re-enable all over again, but no matter how fast I do this again and again, after re-enabling the fc switch port, the failed paths will always go online again after port re-enable
and IO are seen on them. According to the manpage of multipath.conf: "when a path failed event occurs twice in marginal_path_double_failed_time seconds due to an IO error and all the other three parameters are set, multipathd will
fail the path and enqueue this path into a queue...". Did I get anything wrong?
Thanks!
Regards,
Khitai
|
-- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
