[07:23] <jtv2> bigjools: should the test results (maas-test.log) also go to /var/log?  We currently write them to the config/state dir before submitting.
[07:24] <bigjools> jtv2: yes
[07:27] <jtv2> Great.
[07:46] <gmb> Hmm. Importing PXE files seems to have gotten a lot slower overnight... it's now been running for ~30 minutes. That's weird. Oh, mind you, the proxy cache has moved, hasn't it? That might be part of why.
[07:46] <jtv> Hi gmb.
[07:46] <gmb> Hi jtv.
[07:46] <jtv> Yes, it may need to re-initialise.
[07:46] <jtv> Well, to be technical about it, it *will*.
[07:46] <gmb> :)
[07:47] <jtv> But if for whatever reason it wasn't working at all, then this will not present a noticeable cost.
[07:47] <gmb> Right.
[07:48] <gmb> This is my first run of the day — trying to resolve the LP problem. Proxying is _on_ so it should be initialising the cache... I've never seen it take this long, but it's <9am so there's a bit more contention for connections here in sleepytown.
[07:48] <gmb> Once all the kids get off their PS4s it'll speed up.
[07:55] <bigjools> kids playing at earlier than 9am?
[07:55] <bigjools> optimisitic :)
[07:57] <jtv> bigjools: the ones that don't have hangovers can.
[07:57]  * jtv is all faith in kids today
[10:40] <allenap> jtv: Out of interest, why does maas-test have to run as root all the time? Was it just simpler?
[10:42] <jtv> allenap: it may have been the dhcp check.
[10:42] <jtv> Or any place where we need to run our own (or library) python code as root.
[10:44] <jtv> For maintenance, testing & debugging it makes more sense to call that code directly than to externalise it and then call it by shelling out to sudo.
[10:48] <gmb> Running tcpdump *and* strafe to debug maas-import-pxe-files... now I feel like a hacker.
[10:50] <gmb> strafe? No. strace.
[10:53]  * jtv stops googling for this tool new tool gmb was seemingly introducing him to
[10:53] <jtv> cool.
[10:53] <jtv> cool new tool.
[10:53] <jtv> not tool new tool.
[10:53] <rvba> gmb: I think I found the problem.
[10:54] <gmb> rvba: Pray tell.
[10:54] <rvba> The proxy is started with: pidFile=/var/cache/maas-test/proxy.pid
[10:54] <rvba> But then we try to kill it using /run/maas-test/proxy.pid
[10:54] <gmb> Ah, of course.
[10:54] <gmb> Wait, what?
[10:55] <gmb> Huh.
[11:03] <rvba> gmb: in fact, that problem is fixed in trunk.
[11:03] <gmb> Oh, good.
[11:03] <rvba> gmb: if you merge trunk you'll get the fix, you where just unlucky to branch off trunk at the wrong time.
[11:03] <gmb> rvba: Cool, thanks.
[11:25] <rvba> gmb: yeah, probably.
[11:26] <gmb> That does seem to make more sense.
[11:26] <gmb> Unless we want a flood of happy customers.
[11:26] <gmb> :)
[11:26] <rvba> heh
[11:27] <rvba> Still nice to store the report on disk but I don't think we want to encourage people to upload stuff to LP in this case :).
[11:27] <gmb> Yeah :)
[11:28] <gmb> rvba: Okay, I'll make that change shortly.
[11:30] <rvba> gmb: I filed https://bugs.launchpad.net/maas-test/+bug/1267434
[11:30] <gmb> rvba: Okay, thanks.
[11:33] <allenap> jtv: It’s not worth it now, but http://pythonhosted.org/python-prctl/ might have been a better way to do it, rather than going all in for root.
[11:36] <allenap> It’s in universe, not in main though.
[12:07] <gmb> Whoops. Helps if you fix the tests, Graham
[12:10]  * gmb -> lunches
[14:44] <gmb> allenap: Are you free at the top of the hour to talk about the HWE stuff as we discussed this morning?
[14:58] <rvba> gmb: time for a real quick review? https://code.launchpad.net/~rvb/maas-test/get-log/+merge/201020
[14:58] <gmb> rvba: Sure
[14:59] <gmb> rvba: Swapsies? https://code.launchpad.net/~gmb/maas-test/dont-report-successes/+merge/201019
[14:59] <rvba> gmb: deal
[14:59] <gmb> Mine's a bit longer than yours.
[14:59] <gmb> Er...
[14:59] <gmb> I mean...
[14:59] <gmb> never mind.
[14:59] <rvba> :)
[14:59] <gmb> :S
[14:59] <rvba> gmb: btw, I've reviewed your dry-run branch, not sure if you've seen it.
[15:00] <gmb> rvba: I did, thanks; I'll sort that out in a minute.
[15:00] <rvba> cool.
[16:13] <allenap> gmb: I is here.