InfluxDB 1.2.4 leaving .tsm.tmp files, consuming all disk #8526
Comments
|
This should be fixed in the latest nightlies/1.3.0. |
|
Fixed via #8416 |
|
Thankyou for quick reply with positive news! :-) |
|
Not so sure this problem is gone with snapshot (1.4.0~n20170626). Still see a lot of .tsm.tmp files. ` |
|
@forsberg Are you still seeing this issue with 1.3.2 or later? |
|
Well, there's a bunch of them lying around, but there has been no incidents where it's been consuming all the disk during the last ~6 weeks. Current list of files below. Is it normal to see files which are almost 24h old? This is with 1.4.0~n201706260800-0 on amd64. |
|
@forsberg can you grab a goroutine dump on that server using the following command and attach it to this issue |
|
OK, so still on 1.4.0-n201706260800-0: |
|
There's no risk in downgrading that snapshot to 1.3.2? Not sure I'll find the time for that, but maybe. |
|
@forsberg no need to downgrade if you are running master. If that server still has open And then 60 seconds later, grab them all again so we can compare deltas. Very much appreciated |
|
@forsberg, a few more questions:
|
|
@stuartcarnie have we resolved this? |
|
@e-dard yes |
Bug report
I seem to be hitting #7712, although with InfluxDB 1.2.4. So opening new issue as requested.
System info: [Include InfluxDB version, operating system name, and other relevant details]
Steps to reproduce:
Attaching debug info (note: If I try to do the curl provided as example, I just get back "no such profile all") as well as output of 'find /var/lib/influxdb'.
influxdb-debug.ospmon01.1498462213.tar.gz
var-lib-influxdb-files.txt
The text was updated successfully, but these errors were encountered: