This is the second time it happens, it was solved by re-installing couchbase last time, but I think I need to figure out the reason. I only use one small bucket for testing (less than 500 documents), so I don’t think this is related to ram/disk usage. It worked fine until a point when all queries f…

Hi sorry for the late reply, the problem comes again. I’m using CouchBase (Community Edition 5.0.1 build 5003 ) Here’s the google drive link for my indexer.log https://drive.google.com/file/d/1sK1rCNqc9XwrCm31B5sVjrdozwXjOsyu/view?usp=sharing This time, the problem has a different behaviour… Wh…

[image] meowzz: Hi sorry for the late reply, the problem comes again. I’m using CouchBase (Community Edition 5.0.1 build 5003 ) Here’s the google drive link for my indexer.log https://drive.google.com/file/d/1sK1rCNqc9XwrCm31B5sVjrdozwXjOsyu/view?usp=sharing This time, the problem has a diff…

@meowzz , from the logs I see the count of 261 for primary index on POS_Bucket and there is nothing pending to be indexed. I do see this bucket had some deletions. If you have already refreshed the UI and the total count is 270+, we would need the full set of logs to investigate further. You can coll…

@deepkaran.salooja Thanks for your help, Just collected the logs but yesterday while I was poking around, I deleted the primary index and created a index on the field I need to query, then the query works fine again, I do not know if this will affect the logs. Anyway here it is. Google drive link. …

@deepkaran.salooja Another weird thing: select count(*) from BUCKET where FIELD=VALUE return result 19 select * from BUCKET where FIELD=VALUE return result , empty array If there’s no data returned in the second query, shouldn’t it be count=0 in the first query? the log related to this should…

select META().id from BUCKET where FIELD=VALUE; Also post EXPLAIN

@vsr1 Sorry I’ve dropped the primary index and everything works fine again with the field index… I am not able to reproduce the problem at this point…

@deepkaran.salooja Hi, just something that might be related. After I drop the two primary index on two(actually that’s all I have) buckets and create only field index I need, the CPU usage drops to very low. So the CPU usage problem may be related to primary index?

Index on filed can pass the predicates to indexer apply predicates, on primary index predicates are applied post index scan/fetch. Query service has to do lot of work and cpu usage will be more. Please read the articles in http://blog.couchbase.com/wp-content/uploads/2017/03/N1QL-A-Practical-Guide-…

Couchbase N1Q1 query always return empty and dropping-recreating primary index fails

SQL++

deepkaran.salooja May 1, 2018, 11:03pm 2

@meowzz, please share the indexer.log file from /opt/couchbase/var/lib/couchbase/logs. Which couchbase-server version are you using?

Topic		Replies	Views
Could not run any N1QL Couchbase Server	13	2566	September 14, 2016
N1QL which used primary index scan starts working after Re-Creating primary index Couchbase Server	13	1622	September 22, 2017
Created primary index seems ok, but can't find any documents (after some time) SQL++ index	0	1544	March 12, 2017
There is an error related to index Couchbase Server	7	5298	October 17, 2016
Not able to Query SQL++ n1ql	6	6198	December 8, 2015

Couchbase N1Q1 query always return empty and dropping-recreating primary index fails

Related topics