[SOLVED] Upgrade hanging

Fladi

Renowned Member
Feb 27, 2015
31
9
73
Hi,

today I wanted to dist-upgrade my cluster. After the three "VM"-Server I wanted to upgrade the other 3 ceph-severs. But on the first one the "apt-get dist-upgrade" stucked.

[...]
Setting up ceph-common (10.2.6-1~bpo80+1) ...
Setting system user ceph properties..usermod: no changes
..done
Fixing /var/run/ceph ownership....done


After some hour of waiting I killed the process, removed the lock-file and did a "dpkg --configure -a". This now hangs again (same output).

The dpkg.log shows:

2017-03-08 19:11:22 status installed pve-kernel-4.4.40-1-pve:amd64 4.4.40-82
2017-03-08 19:11:22 configure ceph-common:amd64 10.2.6-1~bpo80+1 <none>
2017-03-08 19:11:22 status half-configured ceph-common:amd64 10.2.6-1~bpo80+1


Output of pveversion:

pve-manager/4.4-12/e71b7a74 (running kernel: 4.4.35-2-pve)

What would be the best way to proceed?

Thanks
Tim
 
thanks. found it
"root 7155 7126 0 5622 2488 6 03:23 pts/1 00:00:00 /bin/systemctl start ceph.target"
 
there will be a fixed pve-manager package available soon, which will need to be installed before upgrading the ceph packages (the easiest way is to temporarily comment out the ceph repository in your APT sources). the workaround for hanging upgrades (killing the offending systemctl process) should have no side effects though, so if you have a non-HA setup this also works.
 
(note: only applies if you are using Ceph Jewel and used our upgrade howto or pveceph install!)

you can also apply the (proposed) fix manually before doing the upgrades:
  1. change "WantedBy=ceph.target" to "WantedBy=multi-user.target" in /etc/systemd/system/ceph.service
  2. run "systemctl daemon-reload; systemctl disable ceph.service; systemctl enable ceph.service"
now, "systemctl list-dependencies ceph.target" should no longer list PVE's "ceph.service", and the ceph-common package should no longer trigger a hang on upgrading.
 
  • Like
Reactions: 42n4
Hi,

today I wanted to dist-upgrade my cluster. After the three "VM"-Server I wanted to upgrade the other 3 ceph-severs. But on the first one the "apt-get dist-upgrade" stucked.

[...]
Setting up ceph-common (10.2.6-1~bpo80+1) ...
Setting system user ceph properties..usermod: no changes
..done
Fixing /var/run/ceph ownership....done


After some hour of waiting I killed the process, removed the lock-file and did a "dpkg --configure -a". This now hangs again (same output).

The dpkg.log shows:

2017-03-08 19:11:22 status installed pve-kernel-4.4.40-1-pve:amd64 4.4.40-82
2017-03-08 19:11:22 configure ceph-common:amd64 10.2.6-1~bpo80+1 <none>
2017-03-08 19:11:22 status half-configured ceph-common:amd64 10.2.6-1~bpo80+1


Output of pveversion:

pve-manager/4.4-12/e71b7a74 (running kernel: 4.4.35-2-pve)

What would be the best way to proceed?

Thanks
Tim

executing 'systemctl daemon-reexec' worked for me
 
  • Like
Reactions: compat
Did this as a fix on our HA clusters.
Run the following either before the update or whilst the update has frozen
Code:
sed -i "s/=ceph.target/=multi-user.target/" /etc/systemd/system/ceph.service
systemctl daemon-reload; systemctl disable ceph.service; systemctl enable ceph.service; systemctl daemon-reexec
 
Last edited:

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE, Proxmox Backup Server, and Proxmox Mail Gateway.
We think our community is one of the best thanks to people like you!

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get yours easily in our online shop.

Buy now!