On 5/8/13 10:43 AM, "Klavs Klavsen" <kl...@enableit.dk> wrote:
> >Hi, > >I have 51 clients in puppet-dashboard and when I tell it to delete a >node it simply hangs forever. If I try to just open / on the >puppet-dashboard website I get a proxy timeout (I have apache in front). > >I end up having to kill webrick and start it again with the node NOT >being deleted :( > >I have commented on an issue here - which seems to be about the same: >http://projects.puppetlabs.com/issues/20147#change-90580 > >Anyone have had this problem, and perhaps found a solution? or did you go >in and fired off some sql to delete the node (and which sql) ? > I've been noticing this too. I use the web interface to delete a node and a ruby process spikes up to 100% CPU for quite a long time. Eventually, sometimes approaching 60 minutes, the job finishes and the node is deleted. I've found a more reliable way to delete nodes is to use the the rake job: "rake RAILS_ENV=production node:del name=${FQDN_OF_NODE}". This still takes freaking forever, but doesn't seem to time out. It looks like a fix may be coming eventually based on this request: https://github.com/puppetlabs/puppet-dashboard/pull/226 S.
smime.p7s
Description: S/MIME cryptographic signature