Hi Malte,
so I would only suggest to bring up a new MGR, issue a failover to
that MGR and see if you get the orchestrator to work again.
It should suffice to change the container_image in the unit.run file
(/var/lib/ceph/{FSID}/mgr.{MGR}/unit.run):
CONTAINER_IMAGE={NEWER IMAGE}
So stop one MGR, change the container image, start it and make sure it
takes over as the active MGR.
But I would like to know if I could replace the cephadm on one
running node, stop the MGR and deploy a new MGR on that node with
this:
https://docs.ceph.com/en/latest/cephadm/troubleshooting/#manually-deploying-a-manager-daemon
cephadm --image <container-image> deploy --fsid <fsid> --name
mgr.hostname.smfvfd --config-json config-json.json
This approach probably works as well, but I haven't tried that yet.
And I still do not know what places cephadm... under /var/lib/ceph/fsid.
Does that happen when I enable the orchestrator in the MGR?
And can I replace that cephadm by hand?
The orchestrator would automatically download the respective cephadm
image into that directory if you changed the container_image config
value(s). But I wouldn't do that because you could break your cluster.
If for some reason a MON, OSD or some other Ceph daemon would need to
be redeployed, you would basically upgrade it. That's why I would
suggest to only start one single MGR daemon with a newer version to
see how it goes. In case you get the orchestrator to work again, I
would "downgrade" it again and see what happens next.
Zitat von Eugen Block <eblock@xxxxxx>:
I’m on a mobile phone right now, I can’t go into much detail right now.
But I don’t think it’s necessary to rebuild an entire node, just a
mgr. otherwise you risk cluster integrity if you redeploy a mon as
well with a newer image. I’ll respond later in more detail.
Zitat von Malte Stroem <malte.stroem@xxxxxxxxx>:
Well, thank you, Eugen. That is what I planned to do.
Rebuild the broken node and start a MON and a MGR there with the
latest images. Then I will stop the other MGRs and have a look if
it's working.
But I would like to know if I could replace the cephadm on one
running node, stop the MGR and deploy a new MGR on that node with
this:
https://docs.ceph.com/en/latest/cephadm/troubleshooting/#manually-deploying-a-manager-daemon
cephadm --image <container-image> deploy --fsid <fsid> --name
mgr.hostname.smfvfd --config-json config-json.json
And I still do not know what places cephadm... under /var/lib/ceph/fsid.
Does that happen when I enable the orchestrator in the MGR?
And can I replace that cephadm by hand?
Best,
Malte
On 18.10.24 12:11, Eugen Block wrote:
Okay, then I misinterpreted your former statement:
I think there are entries of the OSDs from the broken node we removed.
So the stack trace in the log points to the osd_remove_queue, but
I don't understand why it's empty. Is there still some OSD removal
going on or something? Did you paste your current cluster status
already? You could probably try starting a Squid mgr daemon by
replacing the container image in the unit.run file and see how
that goes.
Zitat von Malte Stroem <malte.stroem@xxxxxxxxx>:
Hello Eugen,
thanks a lot. However:
ceph config-key get mgr/cephadm/osd_remove_queue
is empty!
Damn.
So should I get a new cephadm with the diff included?
Best,
Malte
On 17.10.24 23:48, Eugen Block wrote:
Save the current output to a file:
ceph config-key get mgr/cephadm/osd_remove_queue > remove_queue.json
Then remove the original_weight key from the json and set the
modified key again with:
ceph config-key set …
Then fail the mgr.
Zitat von Malte Stroem <malte.stroem@xxxxxxxxx>:
Hello Frederic, Hello Eugen,
yes, but I am not sure how to do it.
The links says:
the config-key responsible was mgr/cephadm/osd_remove_queue
This is what it looked like before. After removing the
original_weight field and setting the variable again, the
cephadm module loads and orch works.
So now: Do I remove the value of mgr/cephadm/osd_remove_queue?
Or:
What is meant by:
"After removing the original_weight field and setting the
variable again, the cephadm module loads and orch works."
I can enter a MGR's container and open the file:
/usr/share/ceph/mgr/cephadm/services/osd.py
But what is meant by "removing the original_weight field and
setting the variable again" and what JSON do you mean, Eugen?
osd_obj = OSD.from_json(osd, rm_util=self.rm_util)
Code looks like this:
def load_from_store(self) -> None:
with self.lock:
for k, v in
self.mgr.get_store_prefix('osd_remove_queue').items():
for osd in json.loads(v):
logger.debug(f"Loading osd ->{osd} from store")
osd_obj = OSD.from_json(osd, rm_util=self.rm_util)
if osd_obj is not None:
self.osds.add(osd_obj)
I am a bit lost here.
Best,
Malte
On 17.10.24 21:50, Eugen Block wrote:
I appreciate your kind words. 😎🙂
Frederics link has the correct answer, remove the respective
field from the json.
Zitat von Malte Stroem <malte.stroem@xxxxxxxxx>:
You're so cool, Eugen. Somehow you seem to find out everything.
Yes, this seems to be the issue and I suspected a bug there.
Looking here:
https://github.com/ceph/ceph/blob/main/src/pybind/mgr/cephadm/
services/osd.py
The diff is included in the code.
What can I do now? Get the latest cephadm and put it on the node?
What about the cephadm under /var/lib/ceph/fsid?
I am not sure how to continue.
I would download the latest cephadm and put it under /usr/sbin.
Then disable the module with
ceph mgr module disable cephadm
and enable it
ceph mgr module enable cephadm
Best,
Malte
On 17.10.24 19:20, Eugen Block wrote:
Oh why didn’t you mention earlier that you removed OSDs? 😄
it sounds like this one:
https://tracker.ceph.com/issues/67329
Zitat von Malte Stroem <malte.stroem@xxxxxxxxx>:
Hello Redouane,
thank you. Interesting.
ceph config-key dump
shows about 42000 lines.
What can I search for? Something with OSDs.
But there are thousands of entries.
And if I find something, how can I fix that?
I think there are entries of the OSDs from the broken node
we removed.
Best,
Malte
On 17.10.24 17:46, Redouane Kachach wrote:
So basically it's failing here:
self.to_remove_osds.load_from_store()
This function is responsible of loading Specs from the
mon- store. The
information is stored in json format and it seems the
stored json for the OSD(s) is not valid for some reason.
You can see what's
stored in the mon-store by running:
ceph config-key dump
Don't share the information publicly here especially if it's a
production cluster as it may have sensitive information
about your cluster.
Best,
Redo.
On Thu, Oct 17, 2024 at 5:04 PM Malte Stroem
<malte.stroem@xxxxxxxxx> wrote:
Thanks Eugen & Redouane,
of course I tried enabling and disabling the cephadm
module for the MGRs.
Running ceph mgr module enable cephadm produces this
output in the MGR log:
-1 mgr load Failed to construct class in 'cephadm'
-1 mgr load Traceback (most recent call last):
File "/usr/share/ceph/mgr/cephadm/module.py", line
619, in __init__
self.to_remove_osds.load_from_store()
File "/usr/share/ceph/mgr/cephadm/services/osd.py",
line 922, in
load_from_store
for osd in json.loads(v):
File "/lib64/python3.9/json/__init__.py", line 346, in loads
return _default_decoder.decode(s)
File "/lib64/python3.9/json/decoder.py", line 337, in decode
obj, end = self.raw_decode(s, idx=_w(s, 0).end())
File "/lib64/python3.9/json/decoder.py", line 355, in
raw_decode
raise JSONDecodeError("Expecting value", s,
err.value) from None
json.decoder.JSONDecodeError: Expecting value: line 1
column 1 (char 0)
-1 mgr operator() Failed to run module in active mode
('cephadm')
This comes from inside the MGR container because it's
Python3.9. On the
hosts it'S Python3.11.
I think of redeploying an MGR.
Can I stop the existing MGRs?
Redeploying with ceph orch does not work of course, but I
think this
will work:
https://docs.ceph.com/en/latest/cephadm/troubleshooting/
#manually- deploying-a-manager-daemon
because cephadm standalone is working. Crazy as it sounds.
What do you think?
Best,
Malte
On 17.10.24 12:49, Eugen Block wrote:
Hi,
if you just execute cephadm commands, those are issued
locally on the
hosts, they won't confirm an orchestrator issue immediately.
What does the active MGR log? It could show a stack
trace or error
messages which could point to a root cause.
What about the cephadm files under /var/lib/ceph/fsid?
Can I replace
the latest?
Those are the cephadm versions the orchestrator actually
uses, it will
just download them again from your registry (or upstream).
Can you share:
ceph -s
ceph versions
MGR logs (active MGR)
Thanks,
Eugen
Zitat von Malte Stroem <malte.stroem@xxxxxxxxx>:
Hello,
I am still struggling here and do not know the root
cause of this issue.
Searching the list I found lots of people who had the same or a
similar problem the last years.
However there is no solution four our cluster.
Disabling and enabling the cephadm module does not
work. There are no
error messages. When we run "ceph orch..." we get the
error message:
Error ENOENT: No orchestrator configured (try `ceph
orch set backend`)
But every single cephadm command works!
cephadm ls for example.
Stopping and restarting the MGRs did not help. Removing
the .asok
files did not help.
I think of stopping both MGRs and trying to deploy a
new MGR like this:
https://docs.ceph.com/en/latest/cephadm/troubleshooting/
#manually-
deploying-a-manager-daemon
How could I find the root cause? Is the cephadm somehow broken?
What about the cephadm files under /var/lib/ceph/fsid?
Can I replace
the latest?
Best,
Malte
On 16.10.24 14:54, Malte Stroem wrote:
Hi Laimis,
that did not work. Still ceph orch does not work.
Best,
Malte
On 16.10.24 14:12, Malte Stroem wrote:
Thank you, Laimis.
And you got the same error message? That's strange.
In the mean time I try to check for clients
connected. No Kubernetes
and CephFS, but RGWs.
Best,
Malte
On 16.10.24 14:01, Laimis Juzeliūnas wrote:
Hi Malte,
We have faced this recently when upgrading to Squid
from latest Reef.
As a temporary workaround we disabled the balancer with ‘ceph
balancer off’ and restarted mgr daemons.
We are suspecting older clients (from Kubernetes RBD
mounts as well
as CephFS mounts) on servers with incompatible
client versions but
are yet to dig through it.
Best,
Laimis J.
On 16 Oct 2024, at 14:57, Malte Stroem
<malte.stroem@xxxxxxxxx>
wrote:
Error ENOENT: No orchestrator configured (try `ceph orch set
backend`)
_______________________________________________
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
_______________________________________________
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
_______________________________________________
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
_______________________________________________
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
_______________________________________________
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
_______________________________________________
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
