Nothing but problems with 2.0
Keep getting "Couchbase Server alert: IP address changed" alerts, even after I switch to using hostnames instead of IP addresses.
Now I can't even flush the buckets on one of my nodes, keep getting "Failed to Flush Bucket: Save request failed because of timeout". There's only a few items in this bucket, so no idea why it would timeout.
About ready to revert back to a previous version, 2.0 is not worth the frustration.
For the hostname issue, yes I followed the steps provided in the 'best practices' link you referenced. This did not resolve the issue, still getting the alerts.
Working on diagnostic logs for the flush issue.
How do I send the Diagnostic Log, it's 758 MB in size?
Hello,
I have opened http://www.couchbase.com/issues/browse/MB-7715 for this. It'd be great to post the zipped file to the link above, Thanks!
Anil Kumar
Can you double check if you followed the below two steps and if the issue still remains then I'll open bug to investigate this issue.
* Make sure that this hostname always resolves to the IP address of the node that it is on. This can be accomplished by using a dynamic DNS service such as DNSMadeEasy which will allow you to automatically update the hostname when an underlying IP address changes.
* It is best to make sure that the IP address registered with the hostname is the internal address for the node (rather than the external one provided by Amazon) so that other nodes and application machines can contact it
Not sure I understand your response. "Make sure that this hostname always resolves to the IP address of the node that it is on"??? These servers have fixed IP addreses, they do not change. For some reason, every so often Couchbase is getting confused and thinks the IP address has changed for some reason.
A third question I had was, why do re-balances take so long? I have 2 server nodes with 4 buckets synced between the two. Just as a test, I added one item to one of the buckets, then did a re-balance. It's been running over an hour now. The bucket sizes are only 100MB each.
Thanks,
Matt
file has been attached.
Thanks.
The http://www.couchbase.com/issues/browse/MB-7715 is update with comment from Dev who looked at the logs.
There are many flushes in the log. I took a peek on the one of them only. In that particular case deletion of one of vbuckets took more than 20 seconds. This is known issue. In some cases vbucket deletion can take a lot of time. Sooner or later flush will finish but our UI timeout for this is only 30 seconds. One suggestion is to disable file system barriers or its windows equivalent (see attached image in the bug). It's known to speed up flush significantly.
Hello,
For switching to hostname instead of IP address did you follow the steps provided in 'best practices' (http://www.couchbase.com/docs/couchbase-manual-2.0/couchbase-bestpractic...) section 4.6.2. Handling Changes in IP Addresses.
Regarding the 'Flush Bucket Timeout' issue can you send us the diagnostics logs so that we cab investigate the timeout issue.
Thank you!