[checkmk-commits] [tribe29/checkmk] fecdb0: 13026 FIX agent plugin mk_db2.linux: fix timeout

Branch: refs/heads/1.6.0 Home: https://github.com/tribe29/checkmk Commit: fecdb0f2ee03074d34e1bf459bc586165529ac00 https://github.com/tribe29/checkmk/commit/fecdb0f2ee03074d34e1bf459bc586165… Author: Benedikt Seidl <benedikt.seidl(a)tribe29.com> Date: 2021-07-26 (Mon, 26 Jul 2021) Changed paths: A .werks/13026 M agents/plugins/mk_db2.linux Log Message: ----------- 13026 FIX agent plugin mk_db2.linux: fix timeout When the db2 queries take longer than 10 seconds, the command is terminated. Before this fix the termination did not work out as expected: Only the currently running command was terminated, but the script was still running in the background. This lead to a very confusing output: The output of the terminated script was interleaved with the output of the remaining script. The sections where mixed up and various other db2 checks crashed with strange error messages. The whole problem can be simulated with the following script ran as root: #!/bin/bash function testing { for i in {0..6} do echo $i; sleep 1 done } export -f testing #timeout -s 9 2 su cmkd -s /bin/bash -c "testing" timeout 2 su cmkd -s /bin/bash -c "testing" When running the `-s 9` version, you can see numbers after the timeout occurred, the other version is correctly stopped, as `su` propagates SIGTERM to it's child. Change-Id: I32a1b763f776728e603d2e9fb94bb36e8167e071