0 replica items shown despite replication of 1
I have a default membase bucket with replication 1 in a cluster of 2 servers.
However, in the server tab under replica items there are 2 zeroes.
Additionally, in the vBucket resources tab, there are 1024 vBuckets, but 0 items, and 0.00% replica resident item ratio.
Do I have replication or not?
I rebalanced after adding the second node when initially starting the cluster.
Can you please double check that you have created a membase bucket and not a memcached bucket type ?
Memcached buckets don't support replication
They are both default membase buckets with replication set to 1.....
Replication process for some reason hasent kicked off .
Can you please grab the diagnostics file from :8091/diag or via the link through the web console and send it to us ?
This happened in 2/2 of my 1.7 installs. This is an issue you aren't aware of yet? What is the email address I should send logs to....
Thanks wsorenson.
You can forward them to Farshid@couchbase.com
I will investigate and open an issue in our jira.
It could be related to the environment or firewall or other things that we have not yet seen in our testing
Thanks
Farshid
I'm seeing this same problem. Sometimes replication happens, sometimes it doesn't. (Or the replication is happening but the reporting of the replicated objects isn't happening).
Do you need more logs? Have you any feedback on your investigations so far?
Adam
Adam, we're still investigating this with wsorenson, would it be possible for you to give me SSH access to one of these clusters that is not performing replication?
We have not seen this before (and have not be able to reproduce it internally) but it is obviously a problem that we want to get to the bottom of.
Please sidemail me when you can: perry@couchbase.com
Thanks!
Perry
I believe we've found the problem here. Adam, can you confirm that you're using "flush" to cause this issue?
Perry
I can confirm I am experiencing the same issue. I have used flush via the .net client on our web application. But some of the buckets that were flushed still are replicating/reporting they are replicating on the web console.
I've used the flush mechanism on dev/qa environments but not experienced this before (using 1.7 across all environments and the live environment is obviously getting a lot more throughput).
May be useful information:
I also noted for a little while when looking at a bucket across all server nodes the number of items in the bucket do not match what is reported on the console when looking at number items on individual nodes.
e.g. all server nodes reports for bucketA = ~50K active items, 0 replica items
server node 1 reports for bucketA = ~6K active items, 0 replica items
server node 2 reports for bucketA = ~6K active items, 0 replica items
Thanks guys, this has been fixed already and will be included in 1.7.1 which is due in a few days.
Please let me know if it doesn't solve your issue.
Perry
wsorenson, can you please share more information about your cluster or attach a screenshot of the vbucket-resources ? few questions:1- how many items do have in the bucket ? ( curr_items)2- did you rebalance 2 nodes after adding the second node to the cluster ?