At around 8:58 (PST) this morning, a bunch of talos machines all disconnected from the master, interrupting the current test run: qm-mini-xp02 qm-mini-vista02 qm-mini-vista05 qm-pvista-trunk04 qm-pxp-talos02 qm-pvista-talos02 qm-pxp-talos04 These slaves also disconnected and reconnected, but didn't interrupt builds: qm-plinux-trunk04 qm-mini-vista03

Apparently qm-rhel02 was migrated between VMware hosts at that time: Migrated from bm-vmware10.build.mozilla.org to bm-vmware07.build.mozilla.org by DRS

In bug#467634, I've asked to give qm-rhel02 the same love that was just given to production-master... lets see if thats helps.

I loaded up the buildbot waterfall at around 4:10 (PST) this morning to check on the status. It took a little while to load, and then I see that several clients have just disconnected at 4:12. Could just loading the waterfall be enough to overload the buildbot master and cause disconnections?

And at 4:08 this morning the machine was migrated from bm-vmware09 to bm-vmware11.

I can't think of any action item here, I'm duping this over.