Timing data missing after enabling Time Profile with 1hr timeout #847
Replies: 1 comment
-
|
Please share the output of the workload, notably the section at the end where it gathers the timing. Above in your settings xt_gather_timeout is set to 3600 which adds potential delay of up to an hour to gather all of the timing data, so this should be sufficient. In the medium term Issue #831 is the correct fix for gathering timing data. At present every VU gathers every timed event, this can lead to high use of memory for a large VU count and longer timed tests. After the run every VU passes its data to the monitor VU to sort and process, this is what takes the most time. Instead we need to move to a sampling based approach to limit the amount of data gathered and is likely to resolve issues such as this. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
Hi all,
I’m seeing an issue where HammerDB timing data is missing after enabling Time Profile.
Details:
HammerDB version: 5.0
Benchmark: TPCC
Database: Postgres
Time Profile: Enabled
Test timeout: Set to 1 hr
Issue:
Once Time Profile is enabled and the test is run with a 1hr timeout, the benchmark completes but timing data is not generated or displayed. The run itself appears to execute normally, but the expected timing/profile output is missing.
Any guidance would be appreciated.
Hammerdb cli output :
hammerdb>job 694383D3723B03E213235383 timing 1
No Timing Data for VU 1 for JOB 694383D3723B03E213235383: jobs jobid timing vuid
hammerdb>job 694383D3723B03E213235383 timing 2
No Timing Data for VU 2 for JOB 694383D3723B03E213235383: jobs jobid timing vuid
hammerdb>job 694383D3723B03E213235383 dict
{
"connection": {
"pg_host": "XXXXXXXXX",
"pg_port": "5432",
"pg_sslmode": "prefer"
},
"tpcc": {
"pg_count_ware": "100",
"pg_num_vu": "8",
"pg_superuser": "postgres",
"pg_superuserpass": "postgres_tpcc",
"pg_defaultdbase": "postgres",
"pg_user": "tpcc",
"pg_pass": "tpcc",
"pg_dbase": "benchmarking",
"pg_tspace": "pg_default",
"pg_vacuum": "false",
"pg_dritasnap": "false",
"pg_oracompat": "false",
"pg_cituscompat": "false",
"pg_storedprocs": "true",
"pg_partition": "false",
"pg_total_iterations": "10000000",
"pg_raiseerror": "false",
"pg_keyandthink": "false",
"pg_driver": "timed",
"pg_rampup": "5",
"pg_duration": "60",
"pg_allwarehouse": "false",
"pg_timeprofile": "true",
"pg_async_scale": "false",
"pg_async_client": "10",
"pg_async_verbose": "false",
"pg_async_delay": "1000",
"pg_connect_pool": "false"
}
}
commandline {
sqlite_db = TMP
keepalive_margin = 3600
jobsoutput = JSON
jobs_disable = 0
jobs_profile_id = 0
}
webservice {
ws_port = 8080
}
timeprofile {
profiler = xtprof
xt_unique_log_name = 0
xt_gather_timeout = 3600
xt_job_storage = 1
}
hdb_version {
version = v5.0
}
hammerdb>
Beta Was this translation helpful? Give feedback.
All reactions