Closed mjf closed 1 week ago
Thanks for your feedback. This bug may relate to the merge mode implmenetation in order to be compatible with influxdb protocol. We'll investigate this bug asap.
@mjf Hi, we've release v0.9.3 that includes fix for this bug. Please try this version and check if this bug has been fixed.
@v0y4g3r The crash no longer appears. Fixed, thank you.
What type of bug is this?
Crash
What subsystems are affected?
Standalone mode
Minimal reproduce step
Everything in default after
git pull
andmake
combo (cargo 1.80.1 (376290515 2024-07-16)
):Created database
test
and used Telegraf to ingest some data via[[outputs.influxdb2]]
and using just[[inputs.cpu]]
, collecting every100ms
and flushing data into GreptimeDB every100ms
from 10 threads (10x the same output target inurls
), effectively causing many duplicates.Then run from
psql
:This works.
This causes a crash:
The same holds for
select count(*) from cpu
.When I stop the Telegraf, the crashes no longer occure. If I start the Telegraf again, the described crashes reappear.
What did you expect to see?
No crashes.
What did you see instead?
A crash.
What operating system did you use?
Linux
What version of GreptimeDB did you use?
0.9.2 commit: 8814695b5801f1dadbf6c0bd3f8ffb8864dedd49
Relevant log output and stack trace