Denny [DBA at Innervate]
Denny [DBA at Innervate]
The reason is in Clickhouse server * https://github.com/ClickHouse/ClickHouse/issues/63359
>The target server failed to respond, server ClickHouseNode Client keepalive.timeout=3 / Server keepalive.timeout=3 1 Client successfully executes some query. 2 Client http connection is idle for 2.999999 seconds. 3 Client...
Why this issue is Clickhouse specific? Other databases allows the same postgresql: ``` select 1 result, 2 "RESULT", 3 "ReSuLt"; result | RESULT | ReSuLt --------+--------+-------- 1 | 2 |...
all these settings are available at a client side. ``` clickhouse-client --format=PrettyCompactMonoBlock --output_format_pretty_grid_charset=ASCII --output_format_pretty_color=0 -q "select 1" +-1-+ | 1 | +---+ curl 'http://localhost:8123/?default_format=PrettyCompactMonoBlock&output_format_pretty_grid_charset=ASCII&output_format_pretty_color=0&query=select+1' +-1-+ | 1 | +---+ ```
hm, but only in the case with a single column.
`clickhouse-jdbc-0.4.0-shaded.jar`: ``` SELECT toString(cityHash64(number)) FROM numbers(10000000000) where intDiv(1,number-1000000)>-1000000000000000 java.io.IOException: Reached end of input stream after reading 22 of 51 bytes ```
@strtgbb As I understand you are QA and this is a result of an automation test? Please provide a normal reproducible example (STR), clickhouse-server.log, more information. If you use KILL...
> I call KILL, which does not find anything to kill. I will rerun my test with no KILL at all just in case and report back. Do you mean...
>After every subtest I record the output of I don't follow. Why you need to do it? Why your subtests are not isolated and run against the same table.
> They are not isolated because the intent is to find errors such as those described in this thread. Then your tests must stop if this query return anything `SELECT...