Hi David, Normally cephadm connection issue are not that difficult to solve. It is just the matter of having the appropriate SSH configuration in the root account. Mainly the public key used by cephadm (extracted with the command you used in a shell) added in the root account .ssh/authorized_keys. Normally `ceph -s` should report that there is a cephadm SSH problem and 'ceph health detail' should tell you what are the hosts having a problem. To rerun the test (that runs not very frequently), use ' ceph cephadm check-host host_name". Good luck! Michel Le 02/06/2023 à 20:01, David Barton a écrit :
I am trying to debug an issue with ceph orch host add
Is there a way to debug the specific ssh commands being issued or add debugging code to a python script?
There is nothing useful in my syslog or /var/log/ceph/cephadm.log
Is there a way to get the command to log, or can someone point me in the direction of the source code so I can have a look?
I've run tcpdump on port 22 to listen for outgoing packets and also for traffic going to the target IP, and there is nothing going out when I run ceph orch host add If I run ssh inside the cephadm shell then I see the packets go out and it works as I document below.
I was going to upgrade to Quincy from Pacific 16.2.5 and decided to upgrade from ceph-deploy to cephadm
I initially had problems because I run ssh on a non-standard port. Allowing port 22 has allowed me to run the command below on every node except one.
ceph orch host add [short hostname] [ip address]
That one host fails, inexplicably with the error:
Error EINVAL: Failed to connect to cephstorage-rs01 (103.XXX.YY.ZZ).
If I run cephadm shell (without --no-hosts as that gives the error: unknown flag: --no-hosts) it works as expected.
# cephadm shell Inferring fsid 525ec8aa-b401-4ddf-aa8f-4493727dac02 Inferring config /var/lib/ceph/525ec8aa-b401-4ddf-aa8f-4493727dac02/mon.cephstorage-ig03/config Using recent ceph image ceph/daemon-base@sha256:a038c6dc35064edff40bb7e824783f1bbd325c888e722ec5e814671406216ad5 root@cephstorage-ig03:/# ceph cephadm get-ssh-config > ssh_config root@cephstorage-ig03:/# ceph config-key get mgr/cephadm/ssh_identity_key > ~/cephadm_private_key root@cephstorage-ig03:/# chmod 0600 ~/cephadm_private_key root@cephstorage-ig03:/# ssh -F ssh_config -i ~/cephadm_private_key root@103.XXX.YY.ZZ Warning: Permanently added '103.XXX.YY.ZZ' (ECDSA) to the list of known hosts. Welcome to XXXXX
I was mucking around with custom ssh-config files to get around the port issue, but it did not seem to work so I and reverted back to the vanilla version with: ceph cephadm clear-ssh-config
So when I am inside the shell it works, but it doesn't work properly via ceph orch host add
There is one thing that is unusual that I think is worth mentioning. When I was adding the servers with custom ssh config files, I had a bad entry in the hosts file for cephstorage-rs01 on that server, resolving to 127.0.0.1 When I added it, it said it added the IP as 127.0.0.127# ceph orch host ls
HOST ADDR LABELS STATUS ... cephstorage-rs01 127.0.0.127 Offline ...
I then ran
ceph orch host rm cephstorage-rs01
I have tried an iptables re-route in the vain idea that if there was some kind of host to IP cache it would route to localhost and tell me that the host name didn't match. That did not work.
Right now, I am a sad panda as my ceph cluster is half transitioned. My next port of call is probably to try and adopt what I can into cephadm and make sure the cluster is ok, and then finally drop the problem node and then re-add it.
Any help will be appreciated.
Regards,
David
_______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-leave@ceph.io