Docs Home → Develop Applications → MongoDB Manual
Mitigate Performance Issues with PSA Replica Set
Overview
In a three-member replica set with a primary-secondary-arbiter (PSA) architecture or a sharded cluster with three-member PSA shards, a data-bearing node that is down or lagged can lead to performance issues.
If one data-bearing node goes down, the other node becomes the primary.
Writes with w:1
continue to succeed in this
state but writes with write concern "majority"
cannot
succeed and the commit point starts to lag. If your PSA replica set
contains a lagged secondary and your replica set requires two nodes to
majority commit a change, your commit point also lags.
With a lagged commit point, two things can affect your cluster performance:
The storage engine keeps all changes that happen after the commit point on disk to retain a durable history. The extra I/O from these writes tends to increase over time. This can greatly impact write performance and increase cache pressure.
MongoDB allows the oplog to grow past its configured size limit to avoid deleting the
majority commit point
.
To reduce the cache pressure and increased write traffic, set
votes: 0
and priority: 0
for the node that is unavailable or lagging. For
write operations issued with "majority", only voting members are
considered to determine the number of nodes needed to perform a majority
commit. Setting the configuration of the node to votes: 0
reduces the number of nodes required to commit a
write with write concern "majority"
from two to one and
allows these writes to succeed.
Once the secondary is caught up, you can use the
rs.reconfigForPSASet()
method to set votes
back to 1
.
Note
In earlier versions of MongoDB,
enableMajorityReadConcern
and
--enableMajorityReadConcern
were configurable allowing you
to disable the default read concern "majority"
which
had a similar effect.
Procedure
To reduce the cache pressure and increased write traffic for a
deployment with a three-member primary-secondary-arbiter (PSA)
architecture, set { votes: 0, priority: 0 }
for the secondary that
is unavailable or lagging:
cfg = rs.conf(); cfg["members"][<array_index>]["votes"] = 0; cfg["members"][<array_index>]["priority"] = 0; rs.reconfig(cfg);
If you want to change the configuration of the secondary later, use the
rs.reconfigForPSASet()
method.