starlingx/integ
Code Issues Proposed changes

Adjust Ceph monitoring for multiple monitors on same host

Browse Source 
When locking/unlocking the standby controller, some core dumps were
wrongly generated for the mon.controller process. The problem is in
the Ceph init script when checking for process hung. The regular
expression to filter the ceph health detail command when getting the
status of the monitors was expecting only one monitor per host.

With the new regular expression the script is now correctly checking
the status of each monitor.

Story: 2011122
Task: 50824

Test-Plan:
  PASS: Install AIO-DX and AIO-DX+, lock standby controller and check
  there is no file named hang_trace_mon* in /var/log/ceph directory.

Signed-off-by: Felipe Sanches Zanoni <Felipe.SanchesZanoni@windriver.com>
Change-Id: I23fd0fd92f0639fd81baac7521e95384b8e384b5
This commit is contained in:
Felipe Sanches Zanoni 2024-08-13 10:53:50 -03:00
parent 664794ead1
commit bcfb26840b
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
ceph/ceph/debian/deb_folder/ceph-base.ceph.init
View File

@ -375,7 +375,7 @@ is_process_hung() {
elif [ "$type" = "mon" ]; then
# Get monitor status info
local mon_status=$UP
echo "$CEPH_HEALTH_DETAIL" | grep -q -e "^[[:space:]]*$name.*down"
echo "$CEPH_HEALTH_DETAIL" | grep -q -e "^[[:space:]]*$name[[:space:]].*down"
if [ $? -eq 0 ]; then
mon_status=$DOWN
fi