Fail Over Warning: Rebalance recommended, some data does not have the desired number of replicas

[Lasse Schou]

Hi all,

After changing the number of replicas from 1 to 2 through the unsupported /diag/eval call, proposed by Aliaksey (see below), I keep getting this message (see subject) no matter how many times I rebalance. I've even tried replacing all servers one by one, but the message has stuck. What I did was increasing the replicas to 2. I could finish rebalance after that, so I set the replicas down to one again, after which the rebalance worked fine again. Except from this message:

Fail Over Warning: Rebalance recommended, some data does not have the desired number of replicas!

This was the call I used to increase and descrease replicas.

wget -O- --user=Administrator --password=xxxxx --post-data='ReplicasNum = 2, BucketName = "default", {ok, Config} = ns_bucket:get_bucket(BucketName), Map = [lists:sublist(Chain ++ [undefined, undefined, undefined, undefined], ReplicasNum+1) || Chain <- proplists:get_value(map, Config)], ns_bucket:update_bucket_props(BucketName, [{map, Map}, {num_replicas,ReplicasNum}]).' http://127.0.0.1:8091/diag/eval

Any idea how I get rid of the message?

Lasse

[Aliaksey Kandratsenka]

Most likely this happens because you have 2 nodes and full replication (master + 2 replicas) requires 3.

If not please file a bug. That would be something new.

[Lasse Schou]

Hi. I have 5 nodes. And as I wrote I didn't manage to rebalance with 2 replicas so I switched back to 1 replica using the same call. And after many rebalances I get the mentioned message. Any idea? If not, can you explain how I file a bug?

--
You received this message because you are subscribed to the Google Groups "Couchbase" group.
To unsubscribe from this group and stop receiving emails from it, send an email to couchbase+...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

--
Sent from my phone.

[Aliaksey Kandratsenka]

On Fri, May 24, 2013 at 12:44 PM, Lasse Schou <lasse...@gmail.com> wrote:

Hi. I have 5 nodes. And as I wrote I didn't manage to rebalance with 2 replicas so I switched back to 1 replica using the same call. And after many rebalances I get the mentioned message. Any idea? If not, can you explain how I file a bug?

Please file bug here: http://www.couchbase.com/issues/secure/Dashboard.jspa and assign it to me.

and attach output of cbcollect_info from at least one node.

Particularly I'll need your vbucket map but logs may also hint whats up. cbcollect_info has all of that.

With that information I expect I'll be able to understand whats wrong and how to proceed.

[Tugdual Grall]

Hello Lasse,

Have you filed a bug? if yes could you please put it this thread this will allow the community to follow the discussion.

If not do you still have the issue?

regards

Tug

To unsubscribe from this group and stop receiving emails from it, send an email to couchbase+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Lasse Schou]

Still waiting to get access to the JIRA system. Sent out a request.

2013/5/28 Tugdual Grall <tug...@gmail.com>

To unsubscribe from this group and stop receiving emails from it, send an email to couchbase+...@googlegroups.com.

[Kenny Garland]

I'm also seeing this problem on our cluster.

5 nodes

7 couch buckets

1 memcache bucket

Only 3 of the buckets are using replicas. One has 1 replica, another has 2 and the final one has 3 replicas.

These are all Amazon EC2 m1.medium instances. 4 of the instances are running CB 2.0.0 and one is running 2.0.1

 - Ken

[Aliaksey Kandratsenka]

On Thu, Jun 6, 2013 at 1:37 PM, Kenny Garland <garl...@gmail.com> wrote:

I'm also seeing this problem on our cluster.

5 nodes

7 couch buckets

1 memcache bucket

Only 3 of the buckets are using replicas. One has 1 replica, another has 2 and the final one has 3 replicas.

These are all Amazon EC2 m1.medium instances. 4 of the instances are running CB 2.0.0 and one is running 2.0.1

Yes. See http://www.couchbase.com/issues/browse/MB-8395

It safe to ignore this warning.

[Lasse Schou]

I reduced the bucket with 3 replicas to 2 replicas, and it fixed the issue.

2013/6/6 Aliaksey Kandratsenka <alkond...@gmail.com>

--

[Kenny Garland]

To change the replicas you have to delete the bucket and re-create it, does that mean it is possible something else fixed the issue when the delete occurred?

We have been using 3 replicas on this bucket for a long time and only when I added a new server to the cluster did this error appear.

[Aliaksey Kandratsenka]

On Mon, Jun 10, 2013 at 3:44 AM, Kenny Garland <garl...@gmail.com> wrote:

To change the replicas you have to delete the bucket and re-create it, does that mean it is possible something else fixed the issue when the delete occurred?

We have been using 3 replicas on this bucket for a long time and only when I added a new server to the cluster did this error appear.

Ticket I referred to explains it. This only happens for exactly 5 nodes and exactly 3 replicas.

[Rob Pecor]

I am getting this using 3 nodes, 1 replica.  This is on a fresh install with only the default bucket set to 200MB, using Couchbase Bucket Type.

Using 3.0.2-1603 Enterprise Edition (build-1603)

[Warren Lindsey]

Are you using the rack awareness feature?

--
You received this message because you are subscribed to the Google Groups "Couchbase" group.
To unsubscribe from this group and stop receiving emails from it, send an email to couchbase+...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.