MongoDB 3.6.2 is released

128 views
Skip to first unread message

Kelsey Schubert

unread,
Jan 10, 2018, 5:56:04 PM1/10/18
to mongodb-...@googlegroups.com, mongod...@googlegroups.com, mongodb-dev, mongodb...@googlegroups.com

MongoDB 3.6.2 is out and is ready for production deployment. This release contains only fixes since 3.6.1, and is a recommended upgrade for all 3.6 users.


Fixed in this release:

  • SERVER-31267  CollectionCloner fails if collection is dropped between getMore calls

  • SERVER-31625  The contents of {USER} needs to be escaped when querying for the groups using LDAP server

  • SERVER-31684  QueryPlanKilled (operation exceeded time limit) in $changeStream with updateLookup

  • SERVER-31982  Shard does not call config commit chunk migration command with majority writeConcern nor checks for writeConcern errors.

  • SERVER-32246  PID file permission on v3.6 make it not not monitorable using pid file

  • SERVER-32255  UUIDs may be absent from shard secondary local collections

  • SERVER-32282  Aggregation text search returns text score even if it wasn't requested when targeting multiple shards in a sharded cluster

  • SERVER-32396  mongo shell failed to connect with 3.6 connection string SRV

  • SERVER-32430  DocumentSourceSort sorts array documents incorrectly if there is a non-simple collation

  • SERVER-32529  Requiring replSet for shards breaks Queryable Backup


3.6 Release Notes | All Issues | Downloads


For more information about MongoDB 3.6, read the What’s New in MongoDB 3.6 white paper​

Get up to speed on MongoDB 3.6 with the MongoDB 3.6 Online Course


As always, please let us know of any issues.


-- The MongoDB Team


Clive Strong

unread,
Jan 12, 2018, 8:21:05 AM1/12/18
to mongodb-user
Nice.  Updating production to 3.4.10 this weekend and will schedule QA to be updated to 3.6.2 the following weekend!  Happy days.

Carlos

unread,
Feb 7, 2018, 9:54:03 PM2/7/18
to mongodb-user
My cluster is still on v 3.4... how do I upgrade it to v3.6?
Thanks!

Manoj Gangwar

unread,
Feb 8, 2018, 4:59:32 PM2/8/18
to mongodb-user
Carlos,
You can go through below steps ->
Compatibility Changes in MongoDB 3.6

Java Driver Compatibility

Upgrade Config Servers to Replica Set

Upgrade a Sharded Cluster to 3.6

How to write the configuration is YAML format


0) upgrade the binaries from 3.4.x to 3.6.2 on all boxes but do not restart yet.
link ->
tar -zxvf mongodb-linux-x86_64-3.6.2.tgz
mv mongodb-linux-x86_64-3.6.2 mongo-3.6.2
rm -rf *.tgz

1) Important -> Cluster should have enabled access controls and you have to use the bindIp for this version.

a) security:
   authorization: enabled
   keyFile: <key file path>
b) net:
   bindIp: 127.0.0.1,<that server's local IP>
   port: <port of the process>

c) If you are using old formats for the configuration files then you have to change the contents of the conf files to new 'yaml' format. The old way of writing will not work.

d) all mongod, config, arbiters should be in wiredTiger before you proceed to 3.6.2 IF you have to change any of them, do it first at this step.

e) all config servers must be converted from SCCC -> CSRS. Link is given

2) check on every mongod & config and update featureCompatibilityVersion to 3.4

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

3) Take a backup of Config DB & prepare the scripts for restore as well, in advance->

4) Stop the balancer from any mongos port

5) upgrade the binaries from 3.4.x to 3.6.2 on all boxes but do not restart them, yet.
link to download ->
tar -zxvf mongodb-linux-x86_64-3.6.2.tgz
mv mongodb-linux-x86_64-3.6.2 mongo-3.6.2
rm -rf *.tgz

6) Upgrade the config servers, one by one, secondaries & then primary.(242/ 241/ 240)

first 2 secondaries, then connect 1st config(primary) and do the stepdown,

rs.status()
rs.stepDown()
Now, do the restart this last one, too.


7. upgrade the mongod shards, one by one, (1st shard-> secondary first & then primary, and then complete the 2nd shard similar way)

8) Upgrade the arbiters instances if any, one by one.

9) Upgrade the mongos instances, one by one. 

10) Enable backward-incompatible 3.6 features on all primaries and on mongos

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

11) Re-enable the balancer using the mongos port.

sh.startBalancer()
sh.isBalancerRunning()
sh.getBalancerState()
exit

12) Now, modify Replica Set Protocol Version from pv0 to pv1 and restart all mongod, one by one.
connect to all primaries with mongod port,


cfg = rs.conf();
cfg.protocolVersion=1;
rs.reconfig(cfg);

rs.conf();

13) check if all the processes are up in the cluster. And you are done.

Carlos

unread,
Feb 8, 2018, 6:42:34 PM2/8/18
to mongod...@googlegroups.com
Thanks Manoj, that only applies to instances I fully own and have admin privileges on, correct? My question was in the context of a cloud instance managed by Mongo Atlas.


--
You received this message because you are subscribed to the Google Groups "mongodb-user"
group.
 
For other MongoDB technical support options, see: https://docs.mongodb.com/manual/support/
---
You received this message because you are subscribed to a topic in the Google Groups "mongodb-user" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/mongodb-user/zqcqn-WohzI/unsubscribe.
To unsubscribe from this group and all its topics, send an email to mongodb-user...@googlegroups.com.
To post to this group, send email to mongod...@googlegroups.com.
Visit this group at https://groups.google.com/group/mongodb-user.
To view this discussion on the web visit https://groups.google.com/d/msgid/mongodb-user/fea3fece-4681-4b7f-81fc-d17ade62d8dd%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Manoj Gangwar

unread,
Feb 9, 2018, 8:03:27 PM2/9/18
to mongodb-user
Yes, true. I have stated the steps for a Manual Upgrade process for a Sharded cluster (not Atlas!)

Reply all
Reply to author
Forward
0 new messages