[16:50:00] Is there a way to ensure that if you create three VMs on Cloud VPS that they end up on three different labvirts? [16:54:13] harej: not really, no. [16:54:38] What is the likelihood it will end up happening that way anyway out of random chance? [16:55:56] uncertain. there are a lot of variables in that I think [17:22:21] harej: It's fiarly easy for me to do that by hand as a one-off, on the commandline. [17:22:43] It's unlikely to happen by random chance because there's usually a favorite labvirt that gets lots of new VMs until it crosses a threshold [17:24:35] I created a project proposal to experiment with a PaaS called Flynn. It handles its own load balancing and so I want to experiment with having multiple Flynn servers. I feel like to get the most out of that the three VMs should each be on different servers but I am not sure if that's necessary to get failover benefits. [17:33:01] I wouldn't think it would matter much for testing but if you want me to juggle around the VMs just let me know (I can move them after you create them too) [17:59:15] !log tools.wikibugs restarting wb2-irc as it got killed for excess flood [17:59:17] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.wikibugs/SAL [18:00:32] hm [18:00:42] 9958477 0.31428 wb2-irc tools.wikibu dr 06/06/2018 19:30:41 continuous@tools-exec-1437.too 1 [18:00:54] bd808, is it supposed to be sitting there in 'dr' status? [18:01:03] not familiar with SGE [18:03:34] chasemp, any idea? [18:05:30] ah [18:05:33] it's gone [18:05:37] running jsub [18:06:18] so why did it take so long? [18:26:40] (was in an interview) I have no idea, iirc that means the job has died but SGE failed to clean it up appropriately so it can clear up later if it was a load issue or soemthign I imagine [19:00:01] "dr" is a deletion pending state for grid engine. I don't have any good answer for why it sometimes takes a long time for the job to finish. [19:14:44] bd808, hah okay, thanks