Health Check Solutions
On this page
- Host has decreasing available disk space
- Host has excessive disk utilization
- Host has startup warnings
- Host is unreachable
- MongoDB version outdated
- Replica set has an even number of votes
- Replica set has less than three data-bearing nodes
- Replica set has mixed version nodes
- Replica set has more than one arbiter
- Shared cluster has mixed version nodes
- Too many queued operations
- Too much replication lag
This page lists issues that can be raised by a Cloud Manager health check and provides their solutions.
Host has decreasing available disk space
Cloud Manager considers any disk on any host as needing more disk capacity if it estimates that the disk will be full in two weeks or less.
To remedy this problem, move your database to disk(s) with greater capacity.
Host has excessive disk utilization
Cloud Manager considers any disk on any host as having excessive disk utilization if it is actively storing or retrieving data for a prolonged period of time.
To remedy this problem, move your database to disk(s) with greater throughput.
Host has startup warnings
Limits Startup Warning
Process and user limits with low default values can cause a number of issues in the course of normal MongoDB operation. For further information and recommendations, see UNIX ulimit Settings in the MongoDB Manual.
NUMA Enabled Startup Warning
Running MongoDB on a system with NUMA can cause a number of operational problems, including slow performance for periods of time and high system process usage. For further information and recommendations, see MongoDB and NUMA Hardware in the MongoDB Manual.
Readahead
Please see the readahead information
in this section
of the MongoDB Manual for information and recommendations about the
Readahead
startup warning.
Transparent Huge Pages + Defrag
For information and recommendations about the
Transparent Huge Pages and Defrag
startup warning, see
Disable Transparent Huge Pages (THP).
Host is unreachable
The MongoDB Agent connects to each MongoDB process in your deployment to collect diagnostic data.
If your MongoDB Agent cannot connect to a process, consider the following possible resolutions:
Reason | Resolution |
---|---|
Host no longer exists. | Remove host from Cloud Manager. |
Monitoring cannot reach host. | See
Remedies for a Host Down Alert
for possible resolutions. |
MongoDB version outdated
For MongoDB deployments managed by Cloud Manager, Cloud Manager supports safe automatic upgrade and downgrade operations between releases of MongoDB while maximizing the availability of your deployment. Cloud Manager supports upgrade and downgrade operations for sharded clusters, replica sets, and standalone MongoDB instances.
Configure Available MongoDB Versions describes how to choose which versions of MongoDB are available to Cloud Manager.
If Cloud Manager doesn't manage your deployment, manually change the version of MongoDB. The MongoDB Manual provides upgrade tutorials with each release. For example, see Upgrade MongoDB to 4.2 for upgrading to MongoDB 4.2 from an earlier version.
For managed deployments:
In MongoDB Cloud Manager, go to the Deployment page for your project.
If it is not already displayed, select the organization that contains your desired project from the Organizations menu in the navigation bar.
If it's not already displayed, select your desired project from the Projects menu in the navigation bar.
If the Deployment page is not already displayed, click Deployment in the sidebar.
The Deployment page displays.
Go to the Processes page.
Click the Processes tab for your deployment.
The Processes page displays.
For more information and precautions, see Change the Version of MongoDB.
Replica set has an even number of votes
An even number of voting members in a replica set can lead to election issues in the event of a primary node failure. You should consider adding an additional voting node to your replica sets to ensure an odd number of votes.
You can add an arbiter to your replica set to allow an uneven number of members without the overhead of a member that replicates data.
If your deployment is not managed by Cloud Manager, follow the MongoDB Manual's instructions to manually add an arbiter to your replica set.
For managed deployments:
In MongoDB Cloud Manager, go to the Deployment page for your project.
If it is not already displayed, select the organization that contains your desired project from the Organizations menu in the navigation bar.
If it's not already displayed, select your desired project from the Projects menu in the navigation bar.
If the Deployment page is not already displayed, click Deployment in the sidebar.
The Deployment page displays.
Go to the Processes page.
Click the Processes tab for your deployment.
The Processes page displays.
Replica set has less than three data-bearing nodes
We recommend that your replica set includes at least three data-bearing nodes to ensure high availability. For factors that affect high availability, see the MongoDB Manual's pages on
If your deployment is not managed by Cloud Manager, follow the MongoDB Manual's instructions to manually add a node to your replica set.
For managed deployments:
In MongoDB Cloud Manager, go to the Deployment page for your project.
If it is not already displayed, select the organization that contains your desired project from the Organizations menu in the navigation bar.
If it's not already displayed, select your desired project from the Projects menu in the navigation bar.
If the Deployment page is not already displayed, click Deployment in the sidebar.
The Deployment page displays.
Go to the Processes page.
Click the Processes tab for your deployment.
The Processes page displays.
Replica set has mixed version nodes
Because of potential incompatibilities, it is recommended you upgrade outdated versions of MongoDB instances to the most recent in your cluster.
If your deployment is not managed by Cloud Manager, you will need to manually change the version of MongoDB. The MongoDB Manual provides upgrade tutorials with each release. For example, see Upgrade MongoDB to 4.2 for upgrading to MongoDB 4.2 from an earlier version.
For managed deployments:
In MongoDB Cloud Manager, go to the Deployment page for your project.
If it is not already displayed, select the organization that contains your desired project from the Organizations menu in the navigation bar.
If it's not already displayed, select your desired project from the Projects menu in the navigation bar.
If the Deployment page is not already displayed, click Deployment in the sidebar.
The Deployment page displays.
Go to the Processes page.
Click the Processes tab for your deployment.
The Processes page displays.
For more information and precautions, see Change the Version of MongoDB.
Replica set has more than one arbiter
An arbiter is added to a replica set with an even number of members to add a vote in elections for primary. Arbiters always have exactly one vote, and thus allow replica sets to have an uneven number of members, without the overhead of a member that replicates data. Only one arbiter is required to break election ties.
If your deployment is not managed by Cloud Manager, follow the MongoDB Manual's instructions to manually remove a member from your replica set.
For managed deployments:
In MongoDB Cloud Manager, go to the Deployment page for your project.
If it is not already displayed, select the organization that contains your desired project from the Organizations menu in the navigation bar.
If it's not already displayed, select your desired project from the Projects menu in the navigation bar.
If the Deployment page is not already displayed, click Deployment in the sidebar.
The Deployment page displays.
Go to the Processes page.
Click the Processes tab for your deployment.
The Processes page displays.
For more information on deployment architectures, see Replica Set Deployment Architectures in the MongoDB Manual.
Shared cluster has mixed version nodes
The components of the sharded cluster run different versions of MongoDB.
To avoid compatibility issues, use the same version of MongoDB for all
the mongos
and mongod
processes that make up your sharded cluster.
This includes all the mongod
processes used for the cluster's
config servers and shards.
To change the version of a mongod
or mongos
process,
see Change the Version of MongoDB.
Too many queued operations
Queued operations are operations that are waiting to be processed. This may occur when you have reached your hardware capacity or if you have poorly performing queries.
If you have access to Cloud Manager Premium, you can track long running operations using the Cloud Manager Profiler. To enable the profiler tool in Cloud Manager:
In MongoDB Cloud Manager, go to the Deployment page for your project.
If it is not already displayed, select the organization that contains your desired project from the Organizations menu in the navigation bar.
If it's not already displayed, select your desired project from the Projects menu in the navigation bar.
If the Deployment page is not already displayed, click Deployment in the sidebar.
The Deployment page displays.
Go to the Processes page.
Click the Processes tab for your deployment.
The Processes page displays.
If you don’t have access to Cloud Manager Premium, you can still get access to profiling data for statistics about performance and database operations. To read more about profiling databases, see Profile Databases.
Too much replication lag
Replication lag is a delay between an operation on the primary and the application of that operation from the oplog to the secondary. Replication lag can be a significant issue and can seriously affect MongoDB replica set deployments. Excessive replication lag makes "lagged" members ineligible to quickly become primary and increases the possibility that distributed read operations will be inconsistent.
To learn how to troubleshoot replication lag, please see Check the Replication Lag in the MongoDB Manual.