nova syslog logging to /dev/log race condition
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Glance |
Fix Released
|
Undecided
|
Unassigned | ||
OpenStack Compute (nova) |
Fix Released
|
Medium
|
Unassigned | ||
OpenStack Identity (keystone) |
Fix Released
|
Medium
|
Unassigned | ||
neutron |
Fix Released
|
Medium
|
ratalevolamena | ||
oslo.log |
Fix Released
|
Medium
|
Unassigned |
Bug Description
running nova-api-ec2
running rsyslog
service rsyslog restart ; service nova-api-ec2 restart
nova-api-ec2 consumes up to 100% of the available CPU (or at least a full core) and s not responsive. /var/log/
strace the process
sendto(3, "<142>2013-01-18 20:00:22 24882 INFO nova.service [-] Caught SIGTERM, exiting\0", 77, 0, NULL, 0) = -1 ENOTCONN (Transport endpoint is not connected)
service nova-api-ec2 restart fails as upstart already thinks the process has been terminated.
The only way to recover is to pkill -9 nova-api-ec2 and then restart it with 'service nova-api-ec2 restart'.
The same behavior has been seen in all nova-api services.
Changed in keystone: | |
status: | New → Confirmed |
Changed in nova: | |
status: | New → Confirmed |
Changed in keystone: | |
importance: | Undecided → Medium |
summary: |
- nova syslog logging to /dev/log race condition + nova syslog logging to /dev/log race condition in python 2.6 |
Changed in mos: | |
status: | New → Confirmed |
importance: | Undecided → High |
milestone: | none → 5.1 |
assignee: | nobody → MOS Nova (mos-nova) |
no longer affects: | mos |
affects: | oslo-incubator → oslo.log |
Changed in neutron: | |
importance: | Undecided → Medium |
Changed in neutron: | |
assignee: | nobody → ratalevolamena (chris-techno1307) |
status: | New → Confirmed |
status: | Confirmed → New |
assignee: | ratalevolamena (chris-techno1307) → nobody |
Changed in neutron: | |
assignee: | nobody → ratalevolamena (chris-techno1307) |
Changed in oslo.log: | |
status: | Fix Committed → Fix Released |
Changed in keystone: | |
milestone: | none → mitaka-1 |
Changed in keystone: | |
status: | Fix Committed → Fix Released |
Changed in neutron: | |
status: | Fix Committed → Fix Released |
forgot to mention - this is against folsom but it looks to exist in g as well