What to do when Airwave web interface is down

Aruba Employee
Aruba Employee

Environment Information : Airwave version : 7.7.7


Symptoms : As Symptoms for this issue we can see the httpd daemons restarting. We can see the daemons by executing below command,

# wd

Also, when we tail low_level_ service watcher we would be able to see pound in stopped state and httpd daemon restarting.

To confirm the pound status we can check the status of Pound by executing the below command,

[root@localhost mercury]# service pound status
pound is stopped

Since, pound is in stopped state we can check whether postgresql is running or not. (In this scenario it would be in stopped state).

[root@localhost mercury]# service postgresql status
postmaster is stopped

When we try to restart it, it would fail and would give us the following error message,

[root@localhost mercury]# service postgresql restart
Starting postgresql service:    [Failed]
postmaster dead but pid file exists



Cause : To get a better idea on why postgres is not starting we can check the pgsql log under /var/log

# cd /var/log

# less pgsql

Jan 15 10:24:06 rgppaws001 postgres[29137]: [1-2] LOCATION:  CreateLockFile, miscinit.c:919
Jan 15 10:24:52 rgppaws001 postgres[29256]: [1-1] FATAL:  53100: could not write lock file "/tmp/.s.PGSQL.5432.lock": No space left 
on device
Jan 15 10:24:52 rgppaws001 postgres[29256]: [1-2] LOCATION:  CreateLockFile, miscinit.c:919
Jan 15 10:26:31 rgppaws001 postgres[29532]: [1-1] FATAL:  53100: could not write lock file "/tmp/.s.PGSQL.5432.lock": No space left 
on device

The above message states that database is not able to write any more to /tmp/s.PGSQL.5432.lock since there is no space left in "/tmp"

We can do df -h to confirm the space.


Resolution : Clear up the old files in "/tmp" directory. We can use rm -rf command to clear files. Once we get free space under "/tmp" we can try restarting the postgres service.

[root@localhost log]# service postgresql restart
Stopping postgresql service:                               [  OK  ]
Starting postgresql service:                               [  OK  ]

Now the database would be up, as a result Pound and httpd daemons would be up.

Now we would be able to access the web interface.




Version history
Revision #:
1 of 1
Last update:
‎07-14-2014 01:33 AM
Updated by:
Labels (1)

These 2 processes were stopped, but I did not see the same error logs as are mentioned.

I do have errors in the logs relating to 00000: unexpected pageaddr 69E/76E38000 in log sement blahblah

I'm not sure what files are ok to remove under /tmp, although I would think they would all be ok to remove. Also not sure what to do about it long term.


Search Airheads
Showing results for 
Search instead for 
Did you mean: