Details
Description
Try this 4.6.2 or Pre MB-24429 i.e 5.0.0-2920 to identify root cause and find right solution.
create a new bucket. Lets say onebigjson
-Run the following on your newly created bucket: i.e. 1000 documents each around 2MB
cbworkloadgen -n 10.50.51.11:8091 -j -b onebigjson -i 1000 -s 2000000
-Create the following index:
create index idx_onebig_name on onebigjson(name)
-Then execute the following n1ql query over and over and over.
select * from onebigjson where name is not null;
At one point in time, the cluster will stop responding to your queries. temporary files will start to pile up in the /tmp. And even if you stop executing the N1QL's, cluster never returns back to normal.
all n1ql queries on all bucket in that cluster are affected when the problem appears. The only way to successfully query a bucket via n1ql is to restart the affected bucket by changing the cache metadata. Or restart all nodes one at a time.
Attachments
Issue Links
- relates to
-
MB-24752 cbq-engine does not generating opaque correctly
- Closed