influxdata / influxdata/influxdb
Stat tsmFullCompactionQueue is always 0
- Dominant language
- Rust
- Stars
- 31.7k
- Forks
- 3.7k
- Avg merge
- 13h 37m
- Merged PRs (30d)
- 8
Description
Measurement "influxdb_tsm1_engine"
has field "tsmFullCompactionQueue"
which is always zero.
How to reproduce:
* Install InfluxDB v. 1.7.10 (OSS or Enterprise - the same problem).
* Set compactions limit to low number (say 2).
* Create database with larger number or shards.
o Create retention policy with short shard duration (say 1 day)
o Upload some data spanning multiple shards (say 7 days)
o Upload more data, so that compactions take some time enough for monitoring.
* Delete one series, spanning a number shards, larger than compactions limit.
* Observe
o Field "tsmFullCompactionsActive" reaches configured limit.
o Group by tag "id" to see that once one compaction finishes, another kicks in right away from the queue, until all shards are compacted.
o Observe that tsmFullCompactionQueue was always = 0.
Or just check the code (engine.go) or search for 'TSMFullCompactionsQueue' to see that the integer stat is never assigned and will always have default value of 0.
As a developer, I would expect this to be caught during compile time as a warning.
InfluxDB v. 1.7.10
Contributor guide
Assessment
This issue has not been assessed yet.