Operating the Read (Primary) Balancer

You might be wondering: How can I improve performance in my Ceph cluster? One important data point you can check is the read_balance_score on each of your replicated pools.

This metric, available via ceph osd pool ls detail (see Pools for more details) indicates read performance, or how balanced the primaries are for each replicated pool. In most cases, if a read_balance_score is above 1 (for instance, 1.5), this means that your pool has unbalanced primaries and that you may want to try improving your read performance with the read balancer.

Online Optimization

Enabling

To enable automatic read balancing, you must turn on the balancer module (enabled by default in new clusters) and set the mode to read or upmap-read:

ceph balancer on
ceph balancer mode <read|upmap-read>

Both read and upmap-read mode make use of pg-upmap-primary. In order to use pg-upmap-primary, the cluster cannot have any pre-Reef clients.

If you want to use a different balancer or if you want to make your own custom pg-upmap-primary entries, you might want to turn off the balancer in order to avoid conflict:

ceph balancer off

To allow use of the new feature on an existing cluster, you must restrict the cluster to supporting only Reef (and newer) clients. To do so, run the following command:

ceph osd set-require-min-compat-client reef

This command will fail if any pre-Reef clients or daemons are connected to the monitors. To see which client versions are in use, run the following command:

ceph features

Balancer Module

The balancer module for ceph-mgr will automatically balance the number of primary PGs per OSD if set to read or upmap-read mode. See Balancer Module for more information.

Offline Optimization

Primaries are updated with an offline optimizer that is built into the osdmaptool -- ceph osd cluster map manipulation tool.

Grab the latest copy of your osdmap:
```
ceph osd getmap -o om
```
Run the optimizer:
```
osdmaptool om --read out.txt --read-pool <pool name> [--vstart]
```
It is highly recommended that you run the capacity balancer before running the balancer to ensure optimal results. See Using pg-upmap for details on how to balance capacity in a cluster.
Apply the changes:
```
source out.txt
```
In the above example, the proposed changes are written to the output file out.txt. The commands in this procedure are normal Ceph CLI commands that can be run in order to apply the changes to the cluster.

If you are working in a vstart cluster, you may pass the --vstart parameter as shown above so the CLI commands are formatted with the ./bin/ prefix.

Note that any time the number of pgs changes (for instance, if the pg autoscaler [Autoscaling placement groups] kicks in), you should consider rechecking the scores and rerunning the balancer if needed.

To see some details about what the tool is doing, you can pass --debug-osd 10 to osdmaptool. To see even more details, pass --debug-osd 20 to osdmaptool.

Troubleshooting

Removing pg-upmap-primary mappings

For scenarios where you need to manually remove pg-upmap-primary mappings, Ceph provides the following developer-level commands. These commands should be used with caution, as they directly modify primary PG mappings and can impact read performance (this excludes any data movement).

Note

Users affected by #66867 or #61948 may find these commands useful when dealing with unexpected pg-upmap-primary behavior.

To remove a specific pg-upmap-primary mapping, use:

ceph osd rm-pg-upmap-primary <pgid>

If you need to clear all pg-upmap-primary mappings in your cluster, you may use:

ceph osd rm-pg-upmap-primary-all

Unable to Use Kernel Client

If you are unable to use the kernel client to map RBD images or mount a filesystem while pg-upmap-primary mappings are in your cluster, this is because pg-upmap-primary is not yet supported by the kernel client (as of 2025-09-08).

Follow these steps to confirm this scenario:

Confirm that your cluster contains pg-upmap-primary mappings:

ceph osd dump | grep "pg_upmap_primary"

Check for this error message in the kernel log:

$ dmesg | tail

[73393.901029] libceph: mon2 (1)10.64.24.186:6789 feature set mismatch, my 2f018fb87aa4aafe < server's 2f018fb8faa4aafe, missing 80000000
[73393.901037] libceph: mon2 (1)10.64.24.186:6789 missing required protocol features

Those details confirm that the cluster is using features that the kernel client doesn’t support. Until the kernel client supports pg-upmap-primary, you must remove the mappings to successfully perform mounts. You may do so with the following commands:

If using the balancer module, change the mode back to one that does not use pg-upmap-primary. This prevents additional mappings from being made:

ceph balancer mode upmap

Remove all pg-upmap-primary mappings:

ceph osd rm-pg-upmap-primary-all

Brought to you by the Ceph Foundation

The Ceph Documentation is a community resource funded and hosted by the non-profit Ceph Foundation. If you would like to support this and our other efforts, please consider joining now.