| === jfarschman is now known as MilesDenver | ||
| === liam_ is now known as Guest46061 | ||
| === CyberJacob|Away is now known as CyberJacob | ||
| === CyberJacob is now known as CyberJacob|Away | ||
| caribou | Anything that could explain "Unable to get RPC connection for cluster" from call_cluster when both the region & cluster controller are on the same server ? | 10:22 |
|---|---|---|
| caribou | call_cluster is called from get_all_power_types_from_cluster btw | 10:22 |
| roaksoax | caribou: logs? | 12:40 |
| caribou | roaksoax: yep, let me find a place to put them | 13:14 |
| caribou | roaksoax: http://people.canonical.com/~lbouchard/maaslogs/maas/ | 13:23 |
| chrisg_ | woooo came in to seek help configuring my first maas but didnt need it always nice to see | 14:14 |
| jtv | Yes, and we do like to hear about it — it's the sort of thing you usually never hear. :) | 14:16 |
| rvba | allenap: in your `transaction-views` branch, why do you have this 'decorate' utility that you use instead of calling view = make_retry_view(view)? | 14:32 |
| rvba | allenap: why do we need both methods? | 14:32 |
| allenap | rvba: I can't remember off the top of my head! I'll remove it and see what breaks, if anything. | 14:35 |
| gnuoy | if any work arounds to Bug #1391897 occur to anyone I'd be über grateful. It's only a dev cluster so it doesn't have to be production grade | 14:46 |
| ubot5 | bug 1391897 in MAAS "Internal server error when accessing node info and juju deploy fails" [Critical,Triaged] https://launchpad.net/bugs/1391897 | 14:46 |
| allenap | rvba: With http://paste.ubuntu.com/8965500/ I get the following failures: http://paste.ubuntu.com/8965465/ | 14:47 |
| allenap | rvba: This is wrt transaction-views. | 14:47 |
| === jfarschman is now known as MilesDenver | ||
| rvba | allenap: hum, then I wonder if the hybrid approach isn't hiding a problem… | 15:17 |
| gnuoy | gmb, wrt Bug #1391897 , juju is just proxying the error from MaaS isn't it ? I don't understand why it needs a separate juju bug ? | 15:35 |
| ubot5 | bug 1391897 in MAAS "Internal server error when accessing node info and juju deploy fails" [Critical,Triaged] https://launchpad.net/bugs/1391897 | 15:35 |
| allenap | rvba: What do you think might be hidden? | 15:35 |
| rvba | allenap: I had a suspicion but I was wrong… now I think it would be best to make make_retry_view work with what we use to extract the API views but I guess your approach is okay too… | 15:37 |
| allenap | rvba: If I land it I won't mind if you want to change it to something better. | 15:38 |
| rvba | allenap: heh | 15:38 |
| gmb | gnuoy: Because juju should know better than to put raw HTML in its status output. | 15:54 |
| gnuoy | gmb, agreed! | 15:54 |
| gmb | gnuoy: It'll have got an error of some sort from MAAS, so it should be reacting appropriately. | 15:54 |
| gnuoy | gmb, tip top, thanks for the clarification | 15:54 |
| gmb | gnuoy: You're very welcome, squire. | 15:59 |
| * gmb -> afk for a while; back later | 15:59 | |
| plars | smoser: someone pointed me at what looks to be support for setting up swap in maas with fastpath, are there instructions on how to use that anywhere? I don't see any ui option in the node configuration but perhaps I'm just missing it | 16:47 |
| smoser | plars, it actually should just do it. | 16:47 |
| smoser | if you're using utopic level curtin | 16:47 |
| plars | smoser: we are on the trusty version of maas btw: 1.5.4+bzr2294-0ubuntu1.1 | 16:47 |
| smoser | yeah, does not exist there. | 16:48 |
| plars | smoser: this is a production system :( | 16:48 |
| smoser | well, utopic curtin will get SRU'd to trusty | 16:48 |
| smoser | https://bugs.launchpad.net/ubuntu/+source/curtin/+bug/1386394 | 16:48 |
| ubot5 | Ubuntu bug 1386394 in curtin (Ubuntu Trusty) "SRU: catch 14.04 up with curtin trunk features" [Medium,Fix committed] | 16:48 |
| smoser | but honestly, if you need swap, just set up a swap file | 16:48 |
| plars | smoser: got it, thanks | 16:48 |
| jtv | plars, smoser: the Swapspace package will set up swap files dynamically as needed. | 17:15 |
| plars | jtv: cool, I've never seen that before. Does it work well? | 17:20 |
| jtv | plars: it does for me — but then again I wrote it, back in the day. :) It won't work for sudden large allocations though; it needs to see your memory need and respond to it. | 17:21 |
| plars | jtv: how does it do that? | 17:22 |
| jtv | You configure it along the lines of "I would always like to have 80% of total memory/swap space free" | 17:22 |
| plars | ah, ok | 17:22 |
| jtv | It then watches memory usage and adjusts to meet that target. | 17:22 |
| jtv | There's logic to avoid over-frequent creation/deletion too. | 17:22 |
| plars | jtv: so I guess there's a default minimum threshold at which it starts creating swap? | 17:23 |
| jtv | Yes, it's got a minimum free-space threshold, a maximum, and a cooldown period before it will consider deleting swap files. | 17:24 |
| jtv | There's no real need to configure it; it starts with a small swap file to minimise disruption, just enough to meet your target. Additional swap files get bigger. | 17:25 |
| * gmb returns | 17:40 | |
| jtv | blake_r: gmb just got a hostname humongous-government.maas | 18:25 |
| blake_r | haha | 18:25 |
| blake_r | that's usa | 18:26 |
| roaksoax | jtv: any chance you can look into: https://bugs.launchpad.net/maas/+bug/1391897 | 18:40 |
| ubot5 | Ubuntu bug 1391897 in MAAS "Internal server error when accessing node info and juju deploy fails" [Critical,Triaged] | 18:40 |
| jtv | roaksoax: I'm EOD now, but I can pick it up tomorrow, sure. | 18:41 |
| roaksoax | jtv: great! thanks | 18:41 |
| roaksoax | have a good one | 18:41 |
| jtv | nn | 18:41 |
| === alexpilotti_ is now known as alexpilotti | ||
| === liam_ is now known as Guest39243 | ||
| === jfarschman is now known as MilesDenver | ||
| === menn0_ is now known as menn0 | ||
| === menn0_ is now known as menn0 | ||
| === jfarschman is now known as MilesDenver | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!