#crowbar

ogzy, kirves: i have merged my two pull requests now. feel free to try the iso stuff with a clean unpatched repository :)

tboerger_: great, ty

great, I was testing it now, but it's not that easy when the dev tool tries to revert back to the normal branch :D

kirves: tboerger_i will continue with node installation today :), i had seen some ubuntu stuff at barclamsp installations, i hope i can send some pr this time :)

kirves: --no-switch is the thing you searched before :)

yes, I used that at the end, but it's always easier to work with a clean repo :)

i hope i can start on packaging for opensuse soon

but that will be the case for stoney and not roxy... i think...

tboerger_: what about the dtest1 and dtest2 machines tests failing because of logs that can't be opened?

tboerger_: --no-switch is not working for barclamps/crowbar part, it still asks you commiting but valid for the crowbar main directory

tboerger_: i will look forward for the packaging stuff so that i can convert them to ubuntu

kirves: i think that should already be fixed... the logs dir and crowbar dir is owned by the crowbar user now.

kirves: i didn't check detailly, just erased them from chef interface at the end :)

I'll try that and report back to you :)

ogzy: yeah, but first we've got to put all the package stuff onto the build service ;)

afterwards you are free to port it to ubuntu/debian/centos/rhel :)

for the build service there are submit requests, compareable to pull requests on github.

tboerger_: ok

tboerger_: kirves this is my production.log now, http://dpaste.com/1653358/ i was able to see the dashboard but after i deleted the test nodes from chef interface i am getting 500 page any idea?

tboerger_: seems it was not a good idea to delete them from chef interface

ogzy: no good idea... you need to delete the node and the roles as well.

tboerger_: isn't bluepill supposed to be installed on nodes, also, in Roxy?

I remember being like that in Hydrogen

kirves: i dont know where it is used else

I think it was used to keep chef-client up

The problem is that my nodes now don't keep chef-client up and go offline after a timeout

tboerger_: do you remember what barclamp actually install bluepill on a node? Was it provisioner or deployer?

I'm thinking that maybe it is trying to install the wrong bluepill version

kirves_: https://github.com/crowbar/barclamp-crowbar/pul...

it should be chef/cookbooks/bluepill/recipes/default.rb within barclamp-crowbar i think.

k, I'll have a look at it

Even better, I'll first try with a clean install from the new 'clean' repo

:)

tboerger_: were kirves talking about PROCS CRITICAL: 0 processes with command name 'ruby', args 'chef-client'

tboerger_: this the critical error displayed at the nagioa interface for host

tboerger_: ping

I probably found the problem with bluepill on nodes: the crowbar recipe includes bluepill. The bluepill recipe tries to install activesupport gem but the higher version it can find (4.0.3)

Is there a way to force a version for activesupport gem OR pinpoint the exact location where version 4.0.3 is included and cached'

?

damn, gtg now. I'll look at channel history tomorrow.