Qpid and Pulp live check

[waldirio]

Hello,

This PR will cover issue #69 which is responsible for check the qpid and pulp queues.

[ekohl]

You can consider making this a separate function:

qpid_stat() {
  qpid-stat --ssl-certificate=$CERT -b amqps://localhost:5671 -q | grep -E "(resource_manager.*$SRV_FQDN.*|^  queu|^  =|resource_worker.*$SRV_FQDN.*|resource_manager )"
}

Then in check_with_log you can run qpid_stat | tee -a $LOG and in without.

Another option is to write a log function:

#!/bin/bash

LOG=/tmp/log

write_log() {
	if [[ $LOG_OUTPUT == "Y" ]] ; then
		tee -a $LOG
	else
		cat
	fi
}

echo date | write_log

This then works:

$ ./test.sh 
Wed  5 May 09:33:39 CEST 2021
$ LOG_OUTPUT=Y ./test.sh 
Wed  5 May 09:33:46 CEST 2021
$ cat /tmp/log 
Wed  5 May 09:33:46 CEST 2021

I think the latter is the cleanest result since you'll end up with just check() and write_log(). You may be able to find a better name than write_log, but it's about the principle :)

[waldirio]

Hello @ehokl,
Thanks for the heads up and yea, I agree with you.

Changes applied and also I added something new that I believe to be interesting.

Thank you again.
Waldirio

[wbclark]

+1 to the idea of linking to a single KCS. The KCS can include instructions to download and run this script, basic instructions for how to interpret the output, and link to any BZs which are related.

IMO this script should link to the KCS only, but not the specific BZ. The KCS itself should be the authority on which root causes are possible or likely for the symptoms you might see here.

[waldirio]

@ekohl Thank you for all your insights and ideas. I was working today with @wbclark, we did some follow up and also added a single push.

Please, let me know what you think, and thank you both again!!

teamwork++

[mccun934]

tried this on my Satellite and it worked great. I've been doing this by hand forever, nice to have this in a single script