Time |
Nick |
Message |
10:02 |
sdziallas |
#TOPIC Fedora Sugar Weekly Meeting |
10:03 |
|
bernie, mchua, mtd, SMParrish, tomeu, et al: ping! |
10:03 |
mchua |
here |
10:04 |
sdziallas |
mchua: looks like we don't have a big audience today. anyway... |
10:05 |
mchua |
Yeah... |
10:05 |
|
Let's get started anyhow. |
10:05 |
sdziallas |
nods |
10:05 |
|
#TOPIC Deployments |
10:06 |
|
mchua: that's yours, anyway. :) |
10:06 |
mchua |
Ok. One moment, I'm going to linkdump. |
10:06 |
|
#link http://wiki.sugarlabs.org/go/U[…]:Mchua/SoaS_pilot |
10:06 |
|
#info Looking at a 9-stick, 3-netbook SoaS pilot for my aunt's 1st grade class at Cambridge Friends School |
10:06 |
|
#info objective is to test and document smallest possible development/deployment/QA/support/school-admin/parents/teacher ecosystem test. |
10:06 |
|
#info no hardware or software restrictions; power available, wifi possibly available, $1k hardware budget (Peter has given some netbook suggestions) |
10:06 |
|
#info focus on integrating existing curriculum (social studies unit on 'communities' will be looking at the open source community, etc.) rather than having separate "Sugar time" |
10:07 |
|
#info Mel's role will primarily be QA/documentation; working closely with Lynne May (teacher) |
10:07 |
|
#info Still waiting for +1 from school administration; will be doing lots of announcements as soon as we get that. |
10:07 |
|
#info deployment dates mid-Feb through end of May, exact dates tbd when admin goes +1. |
10:07 |
|
EOF |
10:07 |
|
sdziallas: aaaand that's my update. your turn. |
10:07 |
sdziallas |
mchua: that sounds so awesome! |
10:08 |
mchua |
Essentially, though, this is going to happen so we should just move forward with the planning as if it were a definite thing. |
10:08 |
|
With everything except the publicity bit. |
10:08 |
sdziallas |
sounds totally reasonable to me. |
10:08 |
mchua |
Today I'm interested in figuring out what we need to do in the next 2 weeks, tech-wise, to get ready for this. |
10:08 |
sdziallas |
mchua: from what we discussed last time, getting you a bug tracker and obviously the builds is crucial, right? |
10:08 |
mchua |
nods. |
10:09 |
|
From my side, I know I need to get hardware sometime this month. |
10:09 |
|
#action mchua purchase netbooks for SoaS pilot at CFS |
10:09 |
|
I'll likely just be going with the ones Peter recommended unless there's some other sort of sanity check I should perform on these. |
10:10 |
sdziallas |
mchua: I can ping him to see if he has other news, but I guess it'd be good (unless a local store has them and you could try booting before) |
10:10 |
mchua |
And between when I get them and when the SoaS image is ready, I'll try F12 and F13 on them, get smolt profiles, hardware info, etc. to get ready for testing when the image does arrive. |
10:10 |
sdziallas |
#action sdziallas to ping pbrobinson on netbook recommendations |
10:10 |
mchua |
sdziallas: actually, I think I'm ok, Peter's given me like 3 netbook rec emails :) |
10:11 |
sdziallas |
#info rawhide has started composing again with today, we should get nightlies soon. |
10:11 |
|
mchua: awesome! |
10:11 |
mchua |
the useful thing to find out would be more "so I'm going to have these netbooks, any useful QA/testing I can do (maybe for the Mini SIG, etc) before we slap the SoaS image on them? |
10:11 |
sdziallas |
well, a smolt profile would be great anyway, I guess. |
10:11 |
mchua |
Basically, "do stuff with the underlying distro" (I'll be checking stuff like battery life/power management, too) |
10:11 |
|
Yeah, that's easy. |
10:12 |
sdziallas |
I guess it'd be generally good to hear about your experiences with Linux on those Netbooks. |
10:12 |
|
I got the impression that we're somehow lacking this, but Peter will know more on that. |
10:12 |
mchua |
I'm going to do one on F12, one on F13 rawhide, and one on Blueberry just to cross-compare. |
10:12 |
sdziallas |
Sounds cool! |
10:12 |
mchua |
Ok. Yeah, copious notes will be taken. |
10:13 |
|
Just making sure there's nothing pressing I should do with 'em other than poke around and gather what data I can. |
10:13 |
|
Generally speaking, before the new image is out I should also be getting familiar with Blueberry so that I can more easily catch diffs if something comes up we need to spot. |
10:13 |
|
And seeing if I can figure out a backup solution that works with Blueberry. |
10:14 |
sdziallas |
nods, that'd be good. |
10:14 |
mchua |
Just kinda playing around, we'll see what I can get to. |
10:14 |
sdziallas |
actually, it might be interesting to try something like the following: |
10:14 |
|
* use blueberry on one of the netbooks |
10:14 |
|
* cp ~/.sugar to a usb key |
10:15 |
|
* boot the new snapshot up and overwrite the new ~/.sugar with the old one you saved. |
10:15 |
|
nothing too pressing, either (since it's not a really convinient method for most users). |
10:15 |
satellit |
ssh keys? |
10:16 |
mchua |
sdziallas: That's a great first test for whatever QA test case system we set up. :) |
10:16 |
|
...ah, which I suppose should actually be my first to-do. |
10:16 |
|
#action mchua set up test case management system |
10:17 |
|
I'm not sure what I'm going to do for that yet. Semantic Mediawiki seems like the most promising; I'd like to try that unless there's a better ready-made one out there (things may have changed in the last year since I looked.) |
10:17 |
|
I know that using a bugtracker as a test case management system is asking for pain, at least the way I've seen it done. |
10:18 |
|
EOF |
10:18 |
sdziallas |
mchua: yeah... well, I mean Semantic Mediawiki is nice, from what I experienced. but it's probably an equal pain for using it. |
10:19 |
|
mchua: I don't know, we might want to cover the bug tracker thing here, too (and pull it in before moving to the packaging, if other people drop by, too). |
10:19 |
|
mchua: you know it much better than I do, though. :) |
10:21 |
|
mchua: basically, I guess whatever works best for you should be a go for us on the test case system. |
10:21 |
mchua |
The sooner I get a bugtracker, the sooner I can start filing bugs. |
10:21 |
|
#agreed we're using Semantic Mediawiki for the test case system until/unless it Really Doesn't Work Out |
10:22 |
|
I'll talk with Bernie about setting that up somewhere. |
10:22 |
sdziallas |
(shall we move to that tracker point of the agenda or is there anything else we need to cover wrt the deployment?) |
10:22 |
|
bernie: ping! (^^) |
10:23 |
|
mhm. |
10:23 |
|
#action sdziallas to finally settle bug tracker discussion |
10:24 |
mchua |
What are the things that need to be figured out for that decision to be made? |
10:25 |
sdziallas |
I think we've already narrowed down the use-case pretty much, it's a matter of working out what's possible and especially reasonable. |
10:25 |
|
So the bug tracker would target deployments, devs and technical end users and still stay close to one of the SoaS related projects, so that devs could easily jump in there. |
10:26 |
bernie |
returns from coffee break |
10:26 |
mchua |
Ok - to clarify, is this a bugtracker for all of SoaS (nontrivial migration effort, need buy-in and switch-over from a lot more people), or just for this pilot (in which case anything goes)? |
10:26 |
bernie |
reading |
10:26 |
sdziallas |
mchua: I'd like to have this pilot use whatever will be used by the end of the day. |
10:26 |
|
let's hang on for a second, I bet bernie has some insight on the bug tracker thing, too. ;) |
10:27 |
|
#topic Bug Trackers |
10:27 |
bernie |
sdziallas: I'm not done reading the backlog |
10:27 |
|
mchua: do you already have experience with semantic mediawiki? |
10:28 |
|
mchua: I don't |
10:28 |
mchua |
bernie: I do. |
10:28 |
dogi |
http://me.etin.gs/fedora-olpc/[…]0100121_1002.html |
10:28 |
mchua |
bernie: I've installed it in a sandbox, I've administered the wiki side of it, I've worked with it with reasonable fluency as a user. |
10:28 |
bernie |
mchua: last time I checked, it was quite an invasive extension to mediawiki, with plenty of schema changes (not implying anything bad, just that it may not be super-easy to setup) |
10:28 |
mchua |
bernie: I have *not* installed and maintained it as a responsible sysadmin as part of stable infrastructure, but I don't think we need that for this test anyway. ;) |
10:29 |
|
bernie: I would be perfectly happy with a sandbox (a test machine somewhere, or even a VM) with root that I can chuck things up on for 4-5 months. |
10:29 |
bernie |
mchua: if we agree to install SMW on wiki.sl.o, how long can I procrastinate on this before it starts blocking others? |
10:30 |
|
mchua: dogi could set it up for you... or we could use sunjammer |
10:30 |
mchua |
bernie: Feb 1st, I would say. |
10:30 |
bernie |
mchua: let's use sunjammer to save time... you don't even need root, I 'd say |
10:30 |
|
mchua: ok, I can find the time, by then. |
10:30 |
mchua |
bernie: since SMW is nontrivial and I'm worried about what it might do to load times etc (it's pretty heavy) I dunno if you'd rather go with a sandbox so if things blow up we don't have to panic majorly. |
10:31 |
bernie |
mchua: for the time being, I could mirror our wiki to sunjammer and let you do the work there and then we'll decide if it looks stable |
10:31 |
dogi |
bernie, sure but this should be done in the #treehouse so that is transparent and written |
10:31 |
mchua |
Ok. I can also totally start from scratch. |
10:31 |
|
Basically, bernie, I'm willing to be given root on a VM with no support and be expected to figure the rest out on my own. ;) and anything else I can get on top of that in terms of help, infra, etc. I would consider a happy total bonus. |
10:32 |
|
(I'm cognizant that you're really hosed with infra right now and want to not add to that load if I don't have to.) |
10:32 |
dogi |
+1 mchua |
10:32 |
bernie |
mchua: I'd rather start from a copy of our current MW installation so we can see how it performs in the real case |
10:33 |
|
mchua: I can trust you with root on sunjammer too... but I bet you won't need it for any mediawiki work. |
10:33 |
dogi |
bernie, interesting idea |
10:33 |
bernie |
mchua: our mediawiki has its own group and database user |
10:34 |
|
most web services don't require root at all |
10:34 |
mchua |
bernie: Ok. Yeah, once the plugin has been installed and I have admin on the wiki instance itself, I don't need root at all. |
10:34 |
|
Just mostly needed it for installing the plugin, iirc. |
10:34 |
|
bernie: should we figure this out in http://dev.sugarlabs.org/ticket/400? I think it sounds like we know what we want. |
10:35 |
|
sdziallas: what do you need for a bugtracker decision? |
10:35 |
sdziallas |
notes action item first |
10:35 |
bernie |
mchua: you can manipulate the files too: /srv/www-sugarlabs/wiki-devel/ |
10:35 |
|
mchua: I'm rsyncing our current mediawiki instance there |
10:35 |
sdziallas |
#action mchua bernie and dogi to follow up on semantic mediawiki installation |
10:36 |
|
mchua: basically, we need to make a call. if it involves Fedora, we need an OK there. otherwise, it comes down to saying what we want. |
10:36 |
mchua |
thanks for the action item, sdziallas. |
10:37 |
sdziallas |
:) |
10:37 |
mchua |
sdziallas: Ok - what do we need to know to figure out whether it involves Fedora? |
10:37 |
sdziallas |
mchua: we've basically three options. staying with LP. migrating back to SL. or moving to Fedora Infra. |
10:38 |
dogi |
sdziallas, LP? |
10:38 |
sdziallas |
dogi: Launchpad. |
10:38 |
dogi |
hui |
10:38 |
sdziallas |
dogi: we're still there since we piloted it for SL. |
10:39 |
|
dogi: but the number of people helping out, doing triaging and upstreaming reports has heavily decreased, which is not sustainable anymore. |
10:39 |
|
so I think we should move towards one of our (SoaS') upstream projects. |
10:40 |
dogi |
sdziallas, understand |
10:40 |
sdziallas |
I've heard people arguing that SL and SoaS trackers should probably be seperated. |
10:40 |
|
So since the next SoaS build will be built as part of the Fedora Spin process, it might make sense to move there. |
10:40 |
|
mchua: which answers your question ;) |
10:40 |
mchua |
as a trial round for this release, then reassess? |
10:41 |
sdziallas |
mchua: that might be reasonable. |
10:41 |
mchua |
worksforme. |
10:41 |
|
That would give us Trac, and I'm good with Trac. |
10:41 |
sdziallas |
mchua: well, or BZ. |
10:41 |
mchua |
sdziallas: it sounds like, from your descriptions, that there are downsides to everything except Fedora Infra - any reason *not* to move there? |
10:42 |
|
sdziallas: ....TRAC, PLEASE |
10:42 |
sdziallas |
mchua: hehe ;) |
10:42 |
|
mchua: for example using a fedorahosted trac would mean that it's... at least in the URL Fedora-branded. which might bring Marketing up against us, heh. |
10:44 |
|
(yes, both Fedora Hosted and SL Infra would probably give us Trac) |
10:45 |
mchua |
sdziallas: Is that the only concern with Fedora hosting - branding? |
10:45 |
dogi |
mchua, think so |
10:45 |
sdziallas |
going with Fedora Hosted would give us *another* place, besides RH BZ (which is the common place for packaging reports) and SL Trac. |
10:45 |
mchua |
#link http://dev.sugarlabs.org/ticket/400#comment:2 - SMW ticket |
10:45 |
|
bernie: ^^ for the record |
10:46 |
bernie |
mchua: just got it :) |
10:46 |
sdziallas |
mchua: I know that we can adjust the trac logo in fedorahosted for example, though. |
10:46 |
mchua |
sdziallas: ok, so two things: non-SL-branded URL/tracker, and then redundancy/more places to track. |
10:46 |
bernie |
btw, I'm looking for someone who would like to take over trac from me |
10:46 |
mchua |
sdziallas: it sounds like the "but we'll have Yet Another Tracker" is actually what we want since there are strong arguments for this being separate from Fedora packaging requests and SL development both. |
10:47 |
bernie |
someone who loves dealing with python eggs, long-running application servers, logs full of weird error messages... and so on :-) |
10:47 |
sdziallas |
nods. and using Fedora Hosted would allow us to have it still integrated with our upstream projects. |
10:47 |
|
bernie: LOL :) |
10:48 |
dogi |
:) |
10:48 |
mchua |
bernie: Long walks on the beach? How about long walks on the beach? |
10:48 |
sdziallas |
mchua: and we can still have bugs.soas.org or whatever redirect to such an instance, I guess. |
10:49 |
bernie |
mchua: ok, you can have those, and a pony if you want short rides on the beach... but please take trac over ;-) |
10:49 |
mchua |
bernie: I'll ping you in #sugar about it. :) |
10:49 |
|
sdziallas: Ok. So really, the question is branding. |
10:49 |
|
sdziallas: and that's a "if Sean is OK with it, we are go"? |
10:49 |
bernie |
sdziallas: I think you can redirect, not do a real virtual host. |
10:49 |
|
sdziallas: we had the same problem with launchpad |
10:50 |
|
sdziallas: in general, forges and hosting services don't like projects to use their own brand |
10:50 |
sdziallas |
so we need our own tracker at sugaronastick.{com,org}? ;) |
10:51 |
bernie |
sdziallas: the launchpad people told us it was a technical issue, but it smells very much like a strategical issue to me. |
10:51 |
sdziallas |
nods. |
10:51 |
bernie |
sdziallas: I guess we could share the sugarlabs one if you don't mind. do you *really* want trac? it sucks ass :-) |
10:51 |
sdziallas |
mchua: probably, yes. |
10:52 |
|
bernie: I'm close to "I don't care". |
10:52 |
|
bernie: BZ admittedly sucks. ;) |
10:52 |
bernie |
sdziallas: oh yeah, it sucks more. |
10:52 |
mchua |
I don't care, I just want a bugtracker, any bugtracker, to file bugs in. |
10:52 |
|
And I'd like to have it by the time I get the laptops. |
10:52 |
|
So, Feb 1st. |
10:52 |
dirakx |
preffers Trac ;). |
10:52 |
bernie |
mchua: then let's use the SL one for now... if we move away from trac, we'll certainly migrate all bugs anyway |
10:53 |
|
dirakx: cool. want to maintain it? :-) |
10:53 |
sdziallas |
mchua: and I want one that the devs we rely on would use, too (which blocks Launchpad). |
10:53 |
|
bernie: LOL! |
10:53 |
bernie |
evil |
10:54 |
|
seriously, I promise to maintain trac until we can find a replacement (for me or for trac) |
10:54 |
dirakx |
bernie: hehe,,yeah.. why not. :). |
10:54 |
bernie |
at this time, we got rid of spam and the only big problem left is performance |
10:54 |
dirakx |
i know it's a PITA. |
10:54 |
bernie |
dogi: do you think we could move trac to treehouse? |
10:55 |
|
dirakx: I have an experimental 0.12 installation of trac on sunjammer if you'd like to play with it. I couldn't get authentication to work |
10:55 |
|
dirakx: I wrote to coderanger (Noah), the only trac developer I know, but he did not reply yet. |
10:56 |
tomeu |
the problem I have with projects other that upstream sugar using the same instance is that by default all bugs go to me |
10:56 |
|
and we don't have nobody triaging |
10:56 |
dirakx |
bernie: ok i can hack on a little. |
10:56 |
dogi |
hmm like the idea to have a working VM of trac but remember treehouse is more a playground not the final destination of services |
10:56 |
sdziallas |
#action sdziallas to notify the soas of this discussion |
10:56 |
|
s/soas/soas list |
10:57 |
|
tomeu: I see that point. |
10:57 |
|
tomeu: in the early days, people used to report against SoaS as an affected component, which made it go to me, instead of you. |
10:57 |
mchua |
tomeu: can we set SoaS as a component and have sdziallas be the owner? |
10:57 |
tomeu |
if someone is going to be watching the incoming tickets and take the soas ones from the sugar component, then I'm ok |
10:58 |
mchua |
...basically exactly what you both just said. |
10:58 |
tomeu |
mchua: we already had that, the problem is that people didn't set the component |
10:58 |
mchua |
In practice, I will be the one filing bugs from this pilot. |
10:58 |
|
And the one teaching people how to file bugs for this pilot. |
10:58 |
sdziallas |
mchua: and you know what you're doing :) |
10:58 |
tomeu |
so as long as someone takes care of that, I'm ok |
10:58 |
mchua |
So I can basically guarantee that all SoaS tickets from *this* pilot will be filed correctly. |
10:58 |
sdziallas |
tomeu: I think the issue was they confused "component" and "distro" |
10:58 |
tomeu |
sdziallas: yeah, there's some confusion |
10:58 |
mchua |
I can't guarantee other people from not-my-pilot will file SoaS bugs correctly, though. |
10:59 |
sdziallas |
tomeu: because people filed against sugar in SoaS. |
10:59 |
tomeu |
but I think we can expect that people think that the problem they see is in sugar, even if it's soas not booting up |
10:59 |
bernie |
dirakx: it's on sunjammer, in /srv/www-sugarlabs/bugs-devel/ |
10:59 |
tomeu |
anyway, it's easy to set up a filter for bugmail that catches those instances as they come |
10:59 |
sdziallas |
tomeu: what if we just had SoaS listed as a component and nothing else? |
10:59 |
bernie |
dirakx: let me know if you're interested and I'll add you to group sugartrac so you can make changes |
11:00 |
tomeu |
sdziallas: well, there's the possibility of bugs in sugar that only manifest in soas, right? |
11:00 |
dirakx |
bernie: add me pls. |
11:00 |
tomeu |
not sure how important is that, though |
11:00 |
sdziallas |
tomeu: I'd triage it then. |
11:00 |
|
tomeu: and move it to "sugar" as a component and "Fedora" as distro. |
11:00 |
tomeu |
great! |
11:01 |
bernie |
dirakx: done |
11:01 |
dirakx |
bernie: thx. |
11:01 |
bernie |
mchua: added you to sugarwiki-devel |
11:01 |
sdziallas |
bernie: could you either gimme some privs on trac or take SoaS out of the distro list in SL trac? |
11:02 |
bernie |
sdziallas: sure |
11:02 |
sdziallas |
bernie: awesome, thanks :) |
11:02 |
dogi |
has to go ... ciao |
11:02 |
sdziallas |
ciao dogi! |
11:02 |
bernie |
dirakx: it appears here: http://bugs-devel.sugarlabs.org/ |
11:03 |
sdziallas |
#info tickets in SL should be filed against the "SoaS" component |
11:03 |
mchua |
Ok, so SL trac for this pilot. |
11:03 |
sdziallas |
#action sdziallas will take care of moving those tickets appropriately around |
11:03 |
mchua |
is cool with that. |
11:03 |
sdziallas |
mchua: awesome! :) |
11:04 |
|
#action sdziallas to notify the soas list of bug tracker news |
11:05 |
|
looks like it's time to wrap up, right? |
11:05 |
|
packaging discussions next week? |
11:05 |
bernie |
mchua: no SMW needed any more? |
11:06 |
sdziallas |
I think they'd use both? |
11:06 |
mchua |
bernie: SMW still needed, test case system is separate from ticket tracking. |
11:07 |
|
SMW is going to be for "to make sure SoaS works, run this test, report results here" |
11:07 |
|
and SL trac is "I ran a test, and something broke. Let me report it." |
11:07 |
|
sdziallas: yeah, time to wrap, I have to run |
11:07 |
sdziallas |
mchua: oh, okay... |
11:07 |
|
#action other items deferred to next meeting |
11:08 |
|
thanks everybody for attending! |
11:09 |
|
mchua, your deployment plans rock. and bernie, you rock for still bearing with trac. :) |
11:09 |
|
#endmeeting |