Hi,
a couple of things.
First, is there any specific reason why you're upgrading from .1 to
.2? Why not directly to .15? It seems unnecessary and you're risking
upgrading to a "bad" version (I believe it was 16.2.7) if you're
applying evey minor release. Or why not upgrading to Quincy or Reef
directly?
Second, the error message has changed
(https://github.com/ceph/ceph/pull/41257) from
"could not verify host allowed virtual ips"
to
"does not belong to mon public_network".
I saw this just recently during an upgrade of a cluster I didn't
deploy and it turned out to be a misconfiguration issue (mismatch
between cephadm ssh-user and missing pub key). I recommend to verify
the ssh connections.
ceph cephadm get-ssh-config
ceph cephadm get-user
ceph cephadm get-pub-key
It also could be just a timing issue according to Sage's statement in
above PR:
Oh! I know what the problem is. 1897d1c changed the way we store the
per-host network interface/network info. > On upgrade, cephadm
thinks there are no networks on each host until the device refresh
happens.
How long did you wait?
Zitat von Edouard FAZENDA <e.fazenda@xxxxxxx>:
Dear Ceph Community,
I am in the process of upgrading ceph pacific 16.2.1 to 16.2.2 , I have
followed the documentation :
https://docs.ceph.com/en/pacific/cephadm/upgrade/
My cluster is in Healthy state , but the upgrade is not going forward , as
on the cephadm logs I have the following :
# Ceph -W cephadm
2024-03-06T08:39:11.653447+0000 mgr.rke-sh1-1.qskoyj [INF] Upgrade: Need to
upgrade myself (mgr.rke-sh1-1.qskoyj)
2024-03-06T08:39:12.281386+0000 mgr.rke-sh1-1.qskoyj [INF] Upgrade: Updating
mgr.rke-sh1-2.lxmguj
2024-03-06T08:39:12.286096+0000 mgr.rke-sh1-1.qskoyj [INF] Deploying daemon
mgr.rke-sh1-2.lxmguj on rke-sh1-2
2024-03-06T08:39:19.347877+0000 mgr.rke-sh1-1.qskoyj [INF] Filtered out host
rke-sh1-1: could not verify host allowed virtual ips
2024-03-06T08:39:19.347989+0000 mgr.rke-sh1-1.qskoyj [INF] Filtered out host
rke-sh1-3: could not verify host allowed virtual ips
2024-03-06T08:39:19.366355+0000 mgr.rke-sh1-1.qskoyj [INF] Upgrade: Need to
upgrade myself (mgr.rke-sh1-1.qskoyj)
2024-03-06T08:39:19.965822+0000 mgr.rke-sh1-1.qskoyj [INF] Upgrade: Updating
mgr.rke-sh1-2.lxmguj
2024-03-06T08:39:19.969089+0000 mgr.rke-sh1-1.qskoyj [INF] Deploying daemon
mgr.rke-sh1-2.lxmguj on rke-sh1-2
2024-03-06T08:39:26.961455+0000 mgr.rke-sh1-1.qskoyj [INF] Filtered out host
rke-sh1-1: could not verify host allowed virtual ips
2024-03-06T08:39:26.961502+0000 mgr.rke-sh1-1.qskoyj [INF] Filtered out host
rke-sh1-3: could not verify host allowed virtual ips
2024-03-06T08:39:26.973897+0000 mgr.rke-sh1-1.qskoyj [INF] Upgrade: Need to
upgrade myself (mgr.rke-sh1-1.qskoyj)
2024-03-06T08:39:27.623773+0000 mgr.rke-sh1-1.qskoyj [INF] Upgrade: Updating
mgr.rke-sh1-2.lxmguj
2024-03-06T08:39:27.628115+0000 mgr.rke-sh1-1.qskoyj [INF] Deploying daemon
mgr.rke-sh1-2.lxmguj on rke-sh1-2
My public_network is set :
root@rke-sh1-1:~# ceph config dump | grep public_network
mon advanced public_network
10.10.71.0/24
*
Do you have an idea why I have the following error :
Filtered out host: could not verify host allowed virtual ips
Current state of the upgrade :
# ceph orch upgrade status
{
"target_image":
"docker.io/ceph/ceph@sha256:8cdd8c7dfc7be5865255f0d59c048a1fb8d1335f6923996e
2c2d9439499b5cf2",
"in_progress": true,
"services_complete": [],
"progress": "0/35 ceph daemons upgraded",
"message": "Currently upgrading mgr daemons"
}
progress:
Upgrade to 16.2.2 (24m)
[............................]
Thanks for the help.
Best Regards,
Edouard FAZENDA
Technical Support
Chemin du Curé-Desclouds 2, CH-1226 THONEX +41 (0)22 869 04 40
<https://www.csti.ch/> www.csti.ch
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
