On 14/07/21 1:20 pm, Erwin Bogaard wrote:
Hi,
I just upgraded our cluster to pacific 16.2.5.
As I'm curious what cephfs-top could give for insights, I followed the
steps in the documentation.
After enabling the mgr module "stats":
# ceph mgr module ls
...
"enabled_modules": [
"dashboard",
"iostat",
"restful",
"stats",
"zabbix"
...
I tried the following command:
# ceph fs perf stats
{"version": 1, "global_counters": ["cap_hit", "read_latency",
"write_latency", "metadata_latency", "dentry_lease", "opened_files",
"pinned_icaps", "opened_inodes"], "counters": [], "client_metadata": {},
"global_metrics": {}, "metrics": {"delayed_ranks": []}}
As you can see, this returns no info whatsoever.
Probably there is no IO. Even if you have, the queries would expire - we
have QUERY_EXPIRE_INTERVAL in perf_stats.
> QUERY_EXPIRE_INTERVAL = timedelta(minutes=1)
The same with:
# cephfs-top
cluster ceph does not exist
Could you please use the options: --cluster, --id and --conffile and
check if that connects?
Please send me the mgr logs to check.
The actual cluster name is "ceph".
So I don't understand why "ceph fs perf stats" isn't showing any
information.
Maybe another indicator something isn't ritght:
# ceph fs status
cephfs - 0 clients
======
RANK STATE MDS ACTIVITY DNS INOS DIRS CAPS
...
I see "0 clients". When I take a look in the mgr dashboard, I can actually
see all clients. Which are RHEL 7 & 8 cephfs kernel clients.
There is only 1 mds active, and 1 in standby-replay.
I have multiple pools active, but only 1 fs.
Does anyone have a suggestion where I can take a look enable gathering the
stats?
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
