Details
Critical Description
Repro:
Install travel-sample with primary index
|
create a.sql with following statement 50 times |
select x from `travel-sample` where x =10; |
create b.sql with following statement 50 times (t1 non existent bucket) |
select * from t1;
|
In two windows simultaneously run following commands (cbq must be in the path) and check query.log
|
session 1: cbq -e=127.0.0.1:8093 -u=Administrator -p=password < a.sql |
session 2: cbq -e=127.0.0.1:8093 -u=Administrator -p=password < b.sql |
In query log you see
2020-06-07T12:24:34.266-07:00 [INFO] Pool Get returned travel-sample: no connection pool |
2020-06-07T12:24:34.266-07:00 [INFO] Pool Get returned travel-sample: no connection pool |
Some times you see :
2020-06-07T12:24:34.266-07:00 [ERROR] bulkget exceeded MaxBulkRetries for travel-sample(vbid:46,keys:<ud>[landmark_33345]</ud>) |
When this happens query gets error something like this.
"errors": [
{
"code": 12008,
"msg": "Error performing bulk get operation - cause:
",
"retry": true
}
]
if non existing bucket query not run there are no-errors. non-existing bucket should not impact other queries.
