R3.2 setup_all of centos74 cluster failed unable to start supervisor-database
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Juniper Openstack |
Invalid
|
High
|
Unassigned | ||
R3.2 |
Fix Committed
|
High
|
Santosh Gupta | ||
R4.0 |
Fix Committed
|
High
|
Santosh Gupta | ||
R4.1 |
Fix Committed
|
High
|
Santosh Gupta |
Bug Description
For R3.2 of centos74 cluster, setup_all fails unable to start supervisor-
[root@server3 utils]# service supervisor-database status
● supervisor-
Loaded: loaded (/etc/rc.
Active: active (exited) since Thu 2018-05-17 20:09:20 UTC; 1h 30min ago
Docs: man:systemd-
Process: 12135 ExecStart=
May 17 20:08:59 server3 systemd[1]: Starting SYSV: Supervisord instance for Contrail Database Package...
May 17 20:08:59 server3 supervisor-
May 17 20:08:59 server3 supervisor-
May 17 20:08:59 server3 supervisor-
May 17 20:09:20 server3 systemd[1]: Started SYSV: Supervisord instance for Contrail Database Package.
Hint: Some lines were ellipsized, use -l to show in full.
[root@server3 utils]# echo $?
3
[root@server3 utils]#
Logs uploaded to /cs-shared/bugs/.
Arvind V. debugged the issue and found in 3.2, database packages for rpm don’t create user kafka as part of the post install.
For recovery, we can issue the following:
1) “adduser kafka”
2) chown -R kafka /tmp/kafka-logs
3) chown -R kafka /var/log/kafka
4) restart supervisor
description: | updated |
Had to add Kafka user manually in the testbed because the script that adds the user failed mid way, while provisioning.
Here is the relevant snippet, openstack- database- 3.2.10. 0-75.el7. centos. noa 110/110 contrail- utils/contrail- cassandra- status. py’: No such file or directory openstack- database- 3.2.10. 0-75.el7. centos. noarch) scriptlet failed, exit status 1 openstack- database- 3.2.10. 0-75.el7. centos. noarch
2018-05-17 19:41:16:883163: [root@10.0.0.5] out: Installing : contrail-
2018-05-17 19:41:16:883221: [root@10.0.0.5] out: chown: cannot access ‘/usr/share/
2018-05-17 19:41:16:915448: [root@10.0.0.5] out: warning: %post(contrail-
2018-05-17 19:41:16:915541: [root@10.0.0.5] out: Non-fatal POSTIN scriptlet failure in rpm package contrail-
[root@server3 ~]# ls -lh /usr/share/ contrail- utils/contrail- cassandra- status. py contrail- utils/contrail- cassandra- status. py: No such file or directory
ls: cannot access /usr/share/
however we tested on build 147 of 3.2.10.0 and we found the file, probably Wenquing tested against an older build.
root@sangupta- u14:~# ls -lrt /usr/bin/ contrail- cassandra- * contrail- cassandra- status -> ../share/ contrail- utils/contrail- cassandra- status. py contrail- cassandra- repair -> ../share/ contrail- utils/contrail- cassandra- repair. py
lrwxrwxrwx 1 contrail contrail 52 May 14 01:40 /usr/bin/
lrwxrwxrwx 1 contrail contrail 52 May 14 01:40 /usr/bin/
so closing this.