16:09 <mdeslaur> #startmeeting
16:09 <meetingology> Meeting started Tue Apr 14 16:09:24 2015 UTC.  The chair is mdeslaur. Information about MeetBot at http://wiki.ubuntu.com/meetingology.
16:09 <meetingology> 
16:09 <meetingology> Available commands: action commands idea info link nick
16:09 <mdeslaur> [topic] Apologies
16:09 <mdeslaur> nobody
16:09 <mdeslaur> [topic] Action review
16:09 <mdeslaur> nothing
16:10 <mdeslaur> [topic] Mailing list archive
16:10 <mdeslaur> so it looks like we've responded on-list to MAAS
16:10 <mdeslaur> whoops
16:10 <mdeslaur> to docker
16:10 <mdeslaur> does anyone have anything else they'd like to say about the new docker proposal?
16:10 <pitti> +1 on that, LGTM now
16:11 <pitti> the previous per-series packaging was scary, but that's gone now
16:11 <mdeslaur> yeah, +1 from me as well.
16:12 <pitti> caribou: *prod*
16:12 <caribou> yes, just want to have TB's opinion on an MRE for sosreport
16:13 <caribou> it is a fast moving project with many new inclusions
16:13 <caribou> and not possible to adhere to SRU rules when adding new plugins
16:13 <infinity> caribou: That description is exactly the opposite of what MREs are (usuaully) for.
16:13 <pitti> what does that do, roughly?
16:13 <caribou> so I think that it would be beneficial to have the current upstream version on Ubuntu for all stable rel
16:14 <caribou> collect configuration & logs on running systems for offline analtysis
16:14 <pitti> i. e. a program which shows and manually sends data, or automatic in the background, etc?
16:14 <mdeslaur> it seems to be pretty self-contained and nothing seems to depend on it, so at least there's that
16:15 <caribou> for instance, a new plugin was added for cloud-init which will not be available to any of the stable release
16:15 <caribou> since it is not part of any stable release and will only make it to "W"
16:15 <pitti> it sends it to a location you configure, or to canoincal?
16:15 <caribou> so the possibility to have the same 'recent' version on stable release would be a bonus
16:15 <slangasek> manual
16:15 <pitti> i. e. with SRUs is there a chance that we'd suddenly leak private data which it didn't before on that release?
16:15 <caribou> infinity: I was told  that MRE would be a solution for that
16:15 <slangasek> it's not really an MRE
16:15 <caribou> pitti: manual
16:15 <slangasek> but that's a nomenclature question
16:16 <pitti> TBH I think I need some more details of what that does, how new versions impact stable releases, etc.
16:16 <slangasek> conceptually, as a tool that's used by support to gather information from a customer's system, I think it makes sense to allow it to be updated
16:17 <slangasek> because the extent of the interface from the user is "run this command, get results back from the Canonical support team"
16:17 <caribou> pitti: there is confidential data scrubbing built in, but there is always a chance of bugs around this
16:17 <pitti> caribou: if that's manual configurations, how do me make sure that newer upstream releases work with older configs, and don't suddenly drop config options/information that's sent, or change their format?
16:17 <slangasek> so it seems analogous to me to other exceptions we've made for software where the server interface has changed
16:17 <slangasek> just that in this case the "server interface" is the support team
16:18 <caribou> pitti: but this would affect the dev release in the same way
16:18 <pitti> ah, so it does send data to Caonical, not to the admin's servers
16:18 <caribou> pitti: the tool doesn't send anything
16:19 <caribou> pitti: it produces a tarball to be uploaded "somewhere" by the user
16:19 <pitti> ah, ok
16:19 <caribou> pitti: the only output is a tarball in /tmlp
16:19 <caribou> s/tmlp/tmp
16:19 <infinity> I'm not against it conceptually.
16:19 <pitti> so it's intended for e. g. the Canoincal support team, so it's ok if the format/content changes?
16:19 <infinity> As Martin says, though, are there config files, is migration guaranteed to be sane, etc?
16:20 <pitti> yeah, I'm mostly interested in what this does structurally, and what's the worst thing that can happen
16:20 <pitti> .. if a new upstream version changes format or drops files, etc.
16:20 <caribou> pitti: worst thing is that some collection would be missing (which is the case in the current situation)
16:20 <mdeslaur> perhaps we could better decide if you sent an email to the list with a description of what the tool does, who uses it, what config is uses, what it produces, and the types of changes that have happened in the past?
16:20 <caribou> pitti: this is the current situation with SRU
16:21 <caribou> mdeslaur: that was my intent, but I wanted a first feeling for it
16:21 <caribou> mdeslaur: no point in formally proposing it if the first reaction is totally negative
16:21 <pitti> I'm generally not opposed to SRU exceptions as long as they are done in a safe and sane way
16:22 <mdeslaur> I'm open to the idea, I think this type of tool is something that is definitely worth considering for an exception
16:22 <caribou> pitti: I think that regression issues would be restricted to the output content
16:22 <pitti> ^ agreed; I woudl just like to understand what exactly it is and what the impact is :)
16:22 <pitti> caribou: right, understood
16:22 <caribou> mdeslaur: pitti: Fine, I will send an email with all the details requested
16:23 <pitti> caribou: so I'm trying to find out whether that would break automatic evaluation of the content
16:23 <mdeslaur> caribou: ok, I think we're all open to the idea, and we'll await your post
16:23 <caribou> thanks to the TB, this will help in writing the email
16:23 <caribou> this is all I had
16:23 <pitti> caribou: thanks
16:23 <mdeslaur> thanks caribou
16:24 <mdeslaur> doesn't look like there was anything else to discuss on the list
16:24 <mdeslaur> [topic] Community bugs
16:24 <mdeslaur> None
16:24 <mdeslaur> [topic] Next chair
16:25 <mdeslaur> looks like it's pitti?
16:25 <infinity> or kees. :P
16:25 <pitti> ack
16:25 <slangasek> ;)
16:25 <mdeslaur> ok, so kees if he's still alive, then pitti
16:25 <infinity> Yes, that.
16:25 <mdeslaur> Does anyone have anything else they would like to discuss?
16:26 <pitti> nothing from me
16:26 <slangasek> not I
16:26 <mdeslaur> ok, that's it for today folks
16:26 <mdeslaur> #endmeeting