[SERVER-10967] Adding shard fails due to $CLUSTER database Created: 30/Sep/13  Updated: 20/Nov/13  Resolved: 07/Nov/13

Status: Closed
Project: Core Server
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Bug Priority: Major - P3
Reporter: Dharshan Rangegowda Assignee: Kyle Erf
Resolution: Duplicate Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Amazon AWS, 2.4.1


Issue Links:
Related
related to SERVER-9111 $CLUSTER & $SERVER databases Closed
Operating System: ALL
Participants:

 Description   

Looks like the 2.4.1 servers have a database called "$CLUSTER". When I add a replica set to a shard it fails because a previously added replica set has a "$CLUSTER" db too. Can I go ahead and delete this db to add a second shard?



 Comments   
Comment by Spencer Brody (Inactive) [ 04/Nov/13 ]

This is a known issue in 2.4.1 (SERVER-9111) that was fixed in later version of the 2.4.x series. I highly recommend upgrading to 2.4.8 which should fix this problem.

Dropping the "$CLUSTER" database is safe, and doing so on all shards may (I suspect it will but I haven't tested) workaround the problem temporarily, allowing you to add the shard until you can do the full upgrade to 2.4.8.

Comment by Dharshan Rangegowda [ 04/Nov/13 ]

I don't remember the exact error message but you get it on the second shard because the first shard also has the $Cluster database. Hope that helps.

Comment by Kyle Erf [ 04/Nov/13 ]

Thanks for the description. Do you remember what error you were getting when you were unable to add the shard?

Comment by Dharshan Rangegowda [ 04/Nov/13 ]

Hi Kyle,

Its a fairly simple installation. I just setup a MongoDB shared cluster with 2 shards. Each shard is a 2+1 replica set. I also have 3 config servers and 3 shard routers. I just deleted the $cluster database and proceeded to add the replica set as a shard. For mongo install I used the 2.4.1 RPM package on CentOS.

Comment by Kyle Erf [ 04/Nov/13 ]

Hi Dharshan, sorry for the delayed response to this issue. The solution to your problem is probably not deleting $CLUSTER, which is not an actual database (if you upgrade to a more recent release, in fact, $CLUSTER shouldn't be visible to you).

So that I can better diagnose your problem, could you (1) please describe the layout of your mongodb cluster and (2) please post the exact error messages you are getting?

Generated at Thu Feb 08 03:24:31 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.