Same errors as below on latest ceph / latest Ubuntu LTS / noble when
updating from reef to squid. The same 'ceph -s' that reports all osd's
are 'up' and 'in' also reports all of them are 'unreachable'. I hate it
when that happens. All OSD/mon/mgr hosts are dual stack, but ceph uses
just ip6 with different subnets/interfaces for public & ceph internals.
On 10/1/24 09:43, Sascha Frey wrote:
Hi,
after upgrading our Ceph cluster from 18.2.2 to 19.2.0, I get the
following error messages:
'ceph status' shows HEALTH_ERR 504 osds(s) are not reachable
but, luckily, everything works fine.
I see these related messages in the monitor logs (these messages show up for each OSD):
2024-10-01T16:26:36.185+0200 7ffbd5600640 -1 log_channel(cluster) log [ERR] : osd.101's public address is not in '2001:638:504:2011::/64' subnet
...
2024-10-01T16:34:01.011+0200 7ffbdba00640 -1 unable to convert chosen address to string: 2001:638:504:2011:9:4:1:1
...
IPv6 is enabled in ceph.conf:
ms bind ipv4 = false
ms bind ipv6 = true
public network = 2001:638:504:2011::/64
Is this a bug?
Thanks,
Sascha
_______________________________________________
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
