On restoring vrouter service, haproxy sometimes fails to start for netns SI
Affects | Status | Importance | Assigned to | Milestone | ||
---|---|---|---|---|---|---|
Juniper Openstack | Status tracked in Trunk | |||||
R1.1 |
Invalid
|
High
|
Divakar Dharanalakota | |||
Trunk |
Invalid
|
High
|
Divakar Dharanalakota |
Bug Description
R1.10 Buid 41 (3 config node, 3 compute node) Ubuntu Havana setup
The 3 compute nodes are nodeh4, nodeh5, nodeg29
200.1.1.2 and 200.1.1.3 are two backend vms on nodeh4 and nodeh5 serving a pool sshpool1
Initially haproxy active/passive is set on nodeh5 and nodeg29
I stopped vrouter agent service on nodeh5 and then on nodeg29
Then brought up agent on nodeh5, nodeg29 and restart agent on nodeh4
Now, for sshpool1 , haproxy was not running on nodeh5(nor nodeh4), but was running only on nodeg29
Divakar is aware of this.
Logs will be in http://
On nodeh5's agent introspect, we see this :
Traceback (most recent call last):
File "/usr/bin/
load_
File "/usr/lib/
vrouter_
File "/usr/lib/
netns_
File "/usr/lib/
self.
File "/usr/lib/
self.
File "/usr/lib/
kwargs.
File "/usr/lib/
namespace)
File "/usr/lib/
root_
File "/usr/lib/
raise RuntimeError(m)
RuntimeError:
Command: ['sudo', 'ip', 'netns', 'delete', 'vrouter-
Exit code: 1
Stdout: ''
Stderr: 'Cannot remove /var/run/
root@nodec22:~# nova list
+------
| ID | Name | Status | Task State | Power State | Networks |
+------
| 95019dc8-
| 9d99dd9b-
| 6394e69d-
| 1605b011-
| 2ed3df86-
+------
root@nodec22:~# neutron lb-vip-list
+------
| id | name | address | protocol | admin_state_up | status |
+------
| f30477bb-
| ea60c9ca-
+------
root@nodec22:~# neutron lb-pool-list
+------
| id | name | lb_method | protocol | admin_state_up | status |
+------
| 4e3a7ff5-
| 1f678989-
+------
root@nodec22:~#
root@nodec22:~# neutron lb-pool-show sshpool1
+------
| Field | Value |
+------
| admin_state_up | True |
| description | |
| id | 1f678989-
| lb_method | ROUND_ROBIN |
| members | ea4dc42f-
| | f708ff9e-
| name | sshpool1 |
| protocol | TCP |
| provider | opencontrail |
| status | ACTIVE |
| subnet_id | 4f0ff0bf-
| tenant_id | a6345dd0a98c4cc
| vip_id | f30477bb-
+------
root@nodec22:~#
No haproxy on nodeh4 for sshpool1:
root@nodeh4:~# ps aux |grep haproxy
haproxy 1645 0.0 0.0 29668 2156 ? Ss Sep25 0:18 /usr/sbin/haproxy -f /etc/haproxy/
root 13689 0.0 0.0 8112 920 pts/3 S+ 04:01 0:00 grep --color=auto haproxy
nobody 25857 0.0 0.0 33484 6752 ? Ss 02:32 0:00 haproxy -f /var/lib/
root@nodeh4:~#
No haproxy on nodeh5 for sshpool1:
root@nodeh5:
haproxy 1778 0.0 0.0 28940 960 ? Ss Sep25 0:00 /usr/sbin/haproxy -f /etc/haproxy/
nobody 18936 0.0 0.0 33472 6820 ? Ss 02:47 0:00 haproxy -f /var/lib/
root 21967 0.0 0.0 8108 924 pts/2 S+ 04:02 0:00 grep --color=auto haproxy
root@nodeh5:
root@nodeg29:~# ps aux |grep hapro
haproxy 1592 0.0 0.0 29668 2148 ? Ss Sep25 0:11 /usr/sbin/haproxy -f /etc/haproxy/
root 10954 0.0 0.0 8112 924 pts/4 S+ 03:27 0:00 grep --color=auto hapro
nobody 29364 0.0 0.0 33468 6812 ? Ss 03:03 0:00 haproxy -f /var/lib/
nobody 31900 0.0 0.0 28980 964 ? Ss 03:08 0:00 haproxy -f /var/lib/
root@nodeg29:~#
We need to update the "ip" utility to latest version. The existing "ip" utility version has name space deletion bug. Once updated to latest, this issue will not be seen.
-Divakar