We've slowly started to transition from embedded Barman to the Barman plugin. The migration works seamless when following the migration guide!
We do notice quite a big difference in archival times between both solutions.
When using the embedded Barman, it takes around0.6-0.7 seconds to archive a wal file.
When using the Barman plugin, it takes around 1.6-1.7 seconds to archive a wal file.
for testing I've used the same database with exactly the same configuration, except for the barman plugin.
On both clusters I used pgbench to generate a bunch of wal files.
On the left is the Barman plugin setup, on the right the embedded Barman setup.
Resources on the postgres container and plugin-barman-cloud container are equal as well:
Have I missed something when migrating to the Barman plugin? It just doesn't feel right that with the plugin archiving is about 166% slower.
I'm using release v0.9.0 of the barman plugin and 1.26.1 of the operator.
We've slowly started to transition from embedded Barman to the Barman plugin. The migration works seamless when following the migration guide!
We do notice quite a big difference in archival times between both solutions.
When using the embedded Barman, it takes around0.6-0.7 seconds to archive a wal file.
When using the Barman plugin, it takes around 1.6-1.7 seconds to archive a wal file.
for testing I've used the same database with exactly the same configuration, except for the barman plugin.
On both clusters I used pgbench to generate a bunch of wal files.
On the left is the Barman plugin setup, on the right the embedded Barman setup.
Resources on the postgres container and plugin-barman-cloud container are equal as well:
Have I missed something when migrating to the Barman plugin? It just doesn't feel right that with the plugin archiving is about 166% slower.
I'm using release v0.9.0 of the barman plugin and 1.26.1 of the operator.