Downgrade 3.6 Sharded Cluster to 3.4

Before you attempt any downgrade, familiarize yourself with the contentof this document.

Downgrade Path

Once upgraded to 3.6, if you need to downgrade, we recommend downgrading to the latest patch release of 3.4.

Create Backup

Optional but Recommended. Create a backup of your database.

Considerations

While the downgrade is in progress, you cannot make changes to thecollection metadata. For example, during the downgrade, donot do any of the following:

Prerequisites

Before downgrading the binaries, you must downgrade the featurecompatibility version and remove any 3.6 features incompatible with 3.4 or earlier versions as outlinedbelow. These steps are necessary only iffeatureCompatibilityVersion has ever been set to "3.6".

1. Downgrade Feature Compatibility Version

  • Connect a mongo shell to the mongos instance.

  • Downgrade the featureCompatibilityVersion to "3.4".

  1. db.adminCommand({setFeatureCompatibilityVersion: "3.4"})

The setFeatureCompatibilityVersion command performs writesto an internal system collection and is idempotent. If for any reasonthe command does not complete successfully, retry the command on themongos instance.

To ensure that all members of the sharded cluster reflect the updatedfeatureCompatibilityVersion, connect to each shard replica setmember and each config server replica set member and check thefeatureCompatibilityVersion:

Tip

For a sharded cluster that has access control enabled, to run thefollowing command against a shard replica set member, you mustconnect to the member as a shard local user.

  1. db.adminCommand( { getParameter: 1, featureCompatibilityVersion: 1 } )

All members should return a result that includes:

  1. "featureCompatibilityVersion" : { "version" : "3.4" }

If any member returns a featureCompatibilityVersion that includeseither a version value of "3.6" or a targetVersion field,wait for the member to reflect version "3.4" before proceeding.

For more information on the returned featureCompatibilityVersionvalue, see View FeatureCompatibilityVersion.

2. Remove Backwards Incompatible Persisted Features

Remove all persisted features that are incompatible with 3.4. For example, if you have definedany any view definitions, document validators, and partial indexfilters that use 3.6 query features such as $jsonSchema or$expr, you must remove them.

Procedure

Downgrade a Sharded Cluster

Warning

Before proceeding with the downgrade procedure, ensure that allmembers, including delayed replica set members in the shardedcluster, reflect the prerequisite changes. That is, check thefeatureCompatibilityVersion and the removal of incompatiblefeatures for each node before downgrading.

Download the latest 3.4 binaries.

Using either a package manager or a manual download, get the latestrelease in the 3.4 series. If using a package manager, add a newrepository for the 3.4 binaries, then perform the actual downgradeprocess.

Once upgraded to 3.6, if you need to downgrade, we recommend downgrading to the latest patch release of 3.4.

Disable the Balancer.

Turn off the balancer as described inDisable the Balancer.

Downgrade the mongos instances.

Downgrade the binaries and restart.

Downgrade each shard, one at a time.

Downgrade the shards one at a time. If the shards are replica sets, for each shard:

Note

If you do not perform a clean shut down, errors may result thatprevent the mongod process from starting.

Forcibly terminating the mongod process may causeinaccurate results for db.collection.count() anddb.stats() as well as lengthen startup time the next timethat the mongod process is restarted.

Invoking sudo service mongod stop does not guarantee aclean shutdown. This service script forceably stops themongod process if it takes longer than fiveminutes to shut down.

  • Replace the 3.6 binary with the 3.4 binary.

  • Start the 3.4 binary with the —shardsvr and—port command line options. Include any otherconfiguration as appropriate for your deployment, e.g.—bind_ip.

  1. mongod --shardsvr --port <port> --dbpath <path> \
  2.   --bind_ip localhost,<hostname(s)|ip address(es)>

Or if using a configuration file, update the file toinclude sharding.clusterRole: shardsvr, net.port, and any otherconfiguration as appropriate for your deployment, e.g.net.bindIp, and start:

  1. sharding:
  2.    clusterRole: shardsvr
  3. net:
  4.    port: <port>
  5.    bindIp: localhost,<hostname(s)|ip address(es)>
  6. storage:
  7.    dbpath: <path>
  • Wait for the member to recover to SECONDARY state beforedowngrading the next secondary member. To check the member’sstate, you can issue rs.status() in themongo shell.

Repeat for each secondary member.

  • Step down the replica set primary.

Connect a mongo shell to the primary and users.stepDown() to step down the primary and force anelection of a new primary:

  1. rs.stepDown()

  • When rs.status()shows that the primary has stepped down and another memberhas assumed PRIMARY state, downgrade the stepped-down primary:

    • Shut down the stepped-down primary and replace themongod binary with the 3.4 binary.

    • Start the 3.4 binary with the —shardsvr and—port command line options. Include any otherconfiguration as appropriate for your deployment, e.g.—bind_ip.

  1. mongod --shardsvr --port <port> --dbpath <path>  --bind_ip localhost,<hostname(s)|ip address(es)>

Or if using a configuration file, update the file toinclude sharding.clusterRole: shardsvr, net.port, and any otherconfiguration as appropriate for your deployment, e.g.net.bindIp, and start the 3.4 binary:

  1. sharding:
  2.    clusterRole: shardsvr
  3. net:
  4.    port: <port>
  5.    bindIp: localhost,<hostname(s)|ip address(es)>
  6. storage:
  7.    dbpath: <path>

Downgrade the config servers.

If the config servers are replica sets:

  • Downgrade the secondarymembers of the replica set one at a time:

    • Shut down the secondary mongod instance and replacethe 3.6 binary with the 3.4 binary.

    • Start the 3.4 binary with both the —configsvr and—port options. Include any otherconfiguration as appropriate for your deployment, e.g.—bind_ip.

  1. mongod --configsvr --port <port> --dbpath <path>  --bind_ip localhost,<hostname(s)|ip address(es)>

If using a configuration file, update the file tospecify sharding.clusterRole: configsvr, net.port, and any otherconfiguration as appropriate for your deployment, e.g.net.bindIp, and start the 3.4 binary:

  1. sharding:
  2.    clusterRole: configsvr
  3. net:
  4.    port: <port>
  5.    bindIp: localhost,<hostname(s)|ip address(es)>
  6. storage:
  7.    dbpath: <path>

Include any other configuration as appropriate for your deployment.

  • Wait for the member to recover to SECONDARY state beforedowngrading the next secondary member. To check the member’s state,issue rs.status() in the mongo shell.

Repeat for each secondary member.

  • Step down the replica set primary.

    • Connect a mongo shell to the primary and users.stepDown() to step down the primary and force anelection of a new primary:
  1. rs.stepDown()

  • When rs.status() shows that the primary has steppeddown and another member has assumed PRIMARY state, shut downthe stepped-down primary and replace the mongod binarywith the 3.4 binary.

  • Start the 3.4 binary with both the —configsvr and—port options. Include any otherconfiguration as appropriate for your deployment, e.g.—bind_ip.

  1. mongod --configsvr --port <port> --dbpath <path> --bind_ip localhost,<hostname(s)|ip address(es)>

If using a configuration file, update the file tospecify sharding.clusterRole: configsvr, net.port, and any otherconfiguration as appropriate for your deployment, e.g.net.bindIp, and start the 3.4 binary:

  1. sharding:
  2.    clusterRole: configsvr
  3. net:
  4.    port: <port>
  5.    bindIp: localhost,<hostname(s)|ip address(es)>
  6. storage:
  7.    dbpath: <path>

Re-enable the balancer.

Once the downgrade of sharded cluster components iscomplete, re-enable the balancer.