Upgrading the MDS Cluster

Currently the MDS cluster does not have built-in versioning or file system flags to support seamless upgrades of the MDSs without potentially causing assertions or other faults due to incompatible messages or other functional differences. For this reason, it’s necessary during any cluster upgrade to reduce the number of active MDS for a file system to one first so that two active MDS do not communicate with different versions. Further, it’s also necessary to take standbys offline as any new CompatSet flags will propogate via the MDSMap to all MDS and cause older MDS to suicide.

The proper sequence for upgrading the MDS cluster is:

  1. Reduce the number of ranks to 1:
ceph fs set <fs_name> max_mds 1
  1. Wait for cluster to deactivate non-zero ranks where only rank 0 is active and the rest are standbys.
ceph status # wait for MDS to finish stopping
  1. Take all standbys offline, e.g. using systemctl:
systemctl stop ceph-mds.target
  1. Confirm only one MDS is online and is rank 0 for your FS:
ceph status
  1. Upgrade the single active MDS, e.g. using systemctl:
# use package manager to update cluster
systemctl restart ceph-mds.target
  1. Upgrade/start the standby daemons.
# use package manager to update cluster
systemctl restart ceph-mds.target
  1. Restore the previous max_mds for your cluster:
ceph fs set <fs_name> max_mds <old_max_mds>

Upgrading pre-Firefly filesystems past Jewel


This advice only applies to users with filesystems created using versions of Ceph older than Firefly (0.80). Users creating new filesystems may disregard this advice.

Pre-firefly versions of Ceph used a now-deprecated format for storing CephFS directory objects, called TMAPs. Support for reading these in RADOS will be removed after the Jewel release of Ceph, so for upgrading CephFS users it is important to ensure that any old directory objects have been converted.

After installing Jewel on all your MDS and OSD servers, and restarting the services, run the following command:

cephfs-data-scan tmap_upgrade <metadata pool name>

This only needs to be run once, and it is not necessary to stop any other services while it runs. The command may take some time to execute, as it iterates overall objects in your metadata pool. It is safe to continue using your filesystem as normal while it executes. If the command aborts for any reason, it is safe to simply run it again.

If you are upgrading a pre-Firefly CephFS filesystem to a newer Ceph version than Jewel, you must first upgrade to Jewel and run the tmap_upgrade command before completing your upgrade to the latest version.