Hi metricator team,

we are happily using your App. Thank you for the awesome work.

Sadly, we are running into problems with OracleLinux8 at the moment.

When stopping the SplunkForwarder with systemd, the splunk daemon reaches its default timeout after 360s configured in the system unit file.

When we kill the process with the pid from /opt/splunkforwarder/var/log/metricator/var/fifo_reader_fifo1.pid, the splunkforwarder stops within a few seconds.

We are not sure, but it seems as if the SIGINT signal from systemd to splunk is not propagated down to all subprocesses and therefore splunkd never returns an exit code.

Maybe there is another reason as this only occurs on OL8 in our case.

We’d be happy to assist you in fixing the problem.

Best regards and thank you again

Tested version infos

This issue is reproducible with following component versions:

Splunk Versions:

UF 9.1.1 and 9.1.2
Splunk Enterprise 9.1.2

OS Versions:

Oracle Linux 8 (8.9)
NOT on Oracle Linux 9 or Oracle 7

We further tested different binaries on OL8.9 from /opt/splunkforwarder/var/log/metricator/bin/linux with the very same issue:

./ol/nmon_x86_64_ol8
./generic/nmon_linux_x86_64
./rhel/nmon_x86_64_rhel8

But it seems not to rely on the binary.

‌

Stopping Splunkd or SplunkForwarder using TA-metricator-for-nmon run into TimeoutStopSec=360

Tested version infos

Comments (6)