Bug 747984

Also, this comment, https://bugzilla.redhat.com/show_bug.cgi?id=740688#c6, seems to indicate that if mongod goes down while warehouse is running, warehouse will stay up waiting for mongod to come back online. 

This circumstance needs to be logged as well, something like:
    Oct 21 11:31:11 iwhd[19191]: mongod service in inaccessible
    Oct 21 11:31:11 iwhd[19191]: mongod service in back online

partially addressed (start/stop) by http://git.fedorahosted.org/git?p=iwhd.git;a=commitdiff;h=e24af06c38014e9bda0d4164f64de3f8b51d6a79, which is in iwhd-1.0

Hi,
looks it doesn't work as expected for me against the latest available version on f16 : 

# /bin/iwhd --version
/bin/iwhd (iwhd) 1.0.1-1945

#service iwhd start
log output /var/log/iwhd.log: 
Initializing...
0 replication servers defined
Nov 08 09:21:55 iwhd[14268]: convert-provider: ITER key: name
Nov 08 09:21:55 iwhd[14268]: convert-provider: ITER key: type
Nov 08 09:21:55 iwhd[14268]: convert-provider: ITER key: path

#service iwhd stop
log output /var/log/iwhd.log: 
11 08 09:22:48 killed iwhd[]

If mongod is down then I'm not getting anything in the log.  

Moving bug to ON_DEV status.

I do not understand... The iwhd reports that Mongo is down just fine here:

[root@lembas iwhd-tip]# systemctl start iwhd.service 
[root@lembas iwhd-tip]# tail /var/log/iwhd.log 
Initializing...
0 replication servers defined
Feb 01 17:03:01 iwhd[12039]: convert-provider: ITER key: name
Feb 01 17:03:01 iwhd[12039]: convert-provider: ITER key: type
Feb 01 17:03:01 iwhd[12039]: convert-provider: ITER key: path
Feb 01 17:03:01 iwhd[12039]: server down, no metadata access
[root@lembas iwhd-tip]# 

I noticed some small problems with Jim's fix:
 - [$pid] is empty, because on success pidfile is removed
 - "Initializing" should be log_msg() to get timestamp

I'm going to fix that up, but I think bug is fixed. Dave? Martin?

BTW:
[root@lembas iwhd-tip]# rpm -q iwhd
iwhd-1.2-1.fc16.x86_64

ok, lemme check that with the latest version.

So with iwhd-1.2-3.el6.x86_64 package it still works as is describe in description. When iwhd stops, nothing is logged, when is start then you get 

[root@qeblade ~]# tail /var/log/iwhd.log
Initializing...
0 replication servers defined
Feb 02 06:53:48 iwhd[25988]: convert-provider: ITER key: name
Feb 02 06:53:48 iwhd[25988]: convert-provider: ITER key: type
Feb 02 06:53:48 iwhd[25988]: convert-provider: ITER key: path

if mongod is stop then nothing is logged and output is:

# service mongod status
mongod is stopped
[root@qeblade22 iwhd]# service iwhd status
iwhd is stopped
[root@qeblade22 iwhd]# service iwhd start
waiting for mongod to listen on localhost:27017            [FAILED]
[root@qeblade22 iwhd]# rpm -q iwhd
iwhd-1.2-3.el6.x86_64

Let me know if you need further information.

When mongod is not running, iwhd does not even start, so there is no
way the iwhd binary can log anything.  The init script detects that.
It waits for configurable MONGOD_N_SECONDS (default=2) for the mongod
server to start.  If mongod fails to start in time, the script
prints the "[FAILED]" diagnostic.

The current init script does this:

    wait_for_mongod $MONGOD_N_SECONDS && echo_success \
	|| { echo_failure; echo; return 1; }

it's easy to make it write an explanation to the log:

    wait_for_mongod $MONGOD_N_SECONDS && echo_success \
	|| { echo_failure; echo
	     echo "$(date '+%b %d %T') $PROCESS: failed to start:" \
	       "mongod is inaccessible" >> $IWHD_LOG_FILE
	     return 1; }

IMHO, this isn't worth a release.  I suggest deferring to 1.1.

If someone disagrees, let me know ASAP.

From QA point of view I'm OK with both releases.

Thanks.  Moved to cloudforms-1.1.0