Comment 0 for bug 1626092

Revision history for this message
David Burnazyan (dburnazyan) wrote :

Detailed bug description:
 Syslog
Steps to reproduce:
 1. Create environment with: 3 controller, 20 ceph-osd, 1000 computes
 2. Start deployment
Expected results:
 Logs from all nodes collected on Fuel node
Actual result:
 Logs collected not from all nodes. Syslog use near of 100% of CPU
Reproducibility:
 Each time
Workaround:
 Increase NOFILE limit^
 1. On fuel node create file /etc/systemd/system/rsyslog.service.d/limits.conf with content:
[Service]
LimitNOFILE=16384

  2. execute systemctl daemon-reload
Impact:
 Fuel can't get logs from all nodes
Description of the environment:
 Operation system: Ubuntu-14
 Versions of components: Mitaka
 Reference architecture: 3 controller, 20 ceph-osd, 1000 computes
 Network model: VXLAN
 Related projects installed: none