*** ZogG_laptop has joined #maemo-meeting | 02:53 | |
*** ZogG_laptop has joined #maemo-meeting | 02:53 | |
*** M4rtinK has quit IRC | 04:09 | |
*** lbt has quit IRC | 04:44 | |
*** lbt has joined #maemo-meeting | 04:45 | |
*** DocScrutinizer05 has quit IRC | 06:03 | |
*** DocScrutinizer05 has joined #maemo-meeting | 06:03 | |
*** Pali has joined #maemo-meeting | 10:38 | |
*** Pali has quit IRC | 10:43 | |
*** Pali has joined #maemo-meeting | 10:46 | |
*** M4rtinK has joined #maemo-meeting | 12:53 | |
*** kolp has joined #maemo-meeting | 13:13 | |
*** M4rtinK has quit IRC | 13:55 | |
*** Pali has quit IRC | 15:01 | |
*** Pali has joined #maemo-meeting | 15:01 | |
*** kolp has quit IRC | 15:05 | |
*** kolp has joined #maemo-meeting | 15:12 | |
*** kolp has quit IRC | 15:18 | |
*** Pali has quit IRC | 15:48 | |
*** Pali has joined #maemo-meeting | 15:52 | |
*** kolp has joined #maemo-meeting | 16:52 | |
*** M4rtinK has joined #maemo-meeting | 19:23 | |
*** thedead1440 has joined #maemo-meeting | 19:36 | |
*** thedead1440 has left #maemo-meeting | 19:37 | |
DocScrutinizer05 | [upcoming meeting] alas Ivan/ivgalvez is ill and can't attend (get well soon, pal!), alexander/MentalistTraceur is not able to join due to traveling in no-reception area. Henri/bergie whom I invited to give some details from "inside Nemein" can't join either. | 19:39 |
---|---|---|
kerio | neat | 19:42 |
*** thedead1440 has joined #maemo-meeting | 19:45 | |
*** brkn has joined #maemo-meeting | 19:47 | |
*** mashiara has joined #maemo-meeting | 19:52 | |
Stskeeps | (when does meeting start?) | 19:53 |
*** Woody14619 has joined #maemo-meeting | 19:58 | |
*** Woody14619 has quit IRC | 19:58 | |
*** Woody14619 has joined #maemo-meeting | 19:58 | |
DocScrutinizer05 | [upcoming meeting] alas Ivan/ivgalvez is ill and can't attend (get well soon, pal!), alexander/MentalistTraceur is not able to join due to traveling in no-reception area. Henri/bergie whom I invited to give some details from "inside Nemein" can't join either. | 20:00 |
DocScrutinizer05 | I see mashiara/Rambo/Eero is here. Much appreciated :-) | 20:01 |
DocScrutinizer05 | hi mashiara! | 20:01 |
*** qwazix has joined #maemo-meeting | 20:01 | |
DocScrutinizer05 | hi qwazix! | 20:02 |
qwazix | hi DocScrutinizer05, happy new year! | 20:02 |
DocScrutinizer05 | same to you all! | 20:02 |
DocScrutinizer05 | let me summarize a some old and a few new facts about infra migration | 20:03 |
DocScrutinizer05 | during last few days (3 or so) the autobuilder stopped operation | 20:04 |
DocScrutinizer05 | this is in line with the general read-only notice Eero(?) placed on top of each maemo.org page | 20:04 |
mashiara | that would be X-Fade/Feri | 20:05 |
DocScrutinizer05 | ok, thanks :-) | 20:05 |
DocScrutinizer05 | Generally right ATM Nemein is in a process to move/copy the old servers to interim VM | 20:05 |
DocScrutinizer05 | obviously when you copy something you don't want changes done to source of that copy after it took place | 20:06 |
qwazix | :nod: | 20:06 |
DocScrutinizer05 | there are according messages on maemo community mailinlist | 20:06 |
Woody14619 | Some downtime is to be expected for migration. Just glad it's not off for good. :) | 20:07 |
DocScrutinizer05 | furht er Nemein is installing and setting up 2 dedicated servers for us | 20:07 |
DocScrutinizer05 | (RX300 if you are interested) | 20:07 |
* qwazix is quickly looking up RX300 | 20:08 | |
DocScrutinizer05 | once those machines are up and running, Nemein will migrate the Virtual Machines to those | 20:08 |
DocScrutinizer05 | the main issue of this whole process right now is: we don't have control over the DNS entries pointing to *.maemo.org | 20:09 |
mashiara | um, a moment here | 20:09 |
DocScrutinizer05 | it seems board is in negotiations with Nokia lawyers about transfer of control and responsibility of all maemo related assets | 20:09 |
DocScrutinizer05 | mashiara: ? | 20:10 |
mashiara | the current VMs are not interim, the physical HW was strictly for build-boxes and they will not have enough disk capacity | 20:10 |
mashiara | for running the full suite of services | 20:11 |
DocScrutinizer05 | mhm, so they are meant for parking only? | 20:11 |
mashiara | sorry, can you elaborate what you mean by that ? | 20:11 |
DocScrutinizer05 | the VM are not meant to go operational, but only for temporarily storing the data until moving it to the dedicated servers? | 20:12 |
mashiara | the VMs are meant to go operational, they are in fact ready to go as soon as the DNS issue is sorted | 20:13 |
DocScrutinizer05 | ok, thanks for clarifying it | 20:13 |
qwazix | AIUI they're not going to be hosted on the RX-300's Nokia bought for us | 20:13 |
qwazix | but on other machines | 20:13 |
DocScrutinizer05 | could interested developers in a pinch already use them by addressing them via IP rather than URL? | 20:13 |
DocScrutinizer05 | qwazix: obviously those VM are not hosted on the yet non-existing RX300 | 20:14 |
mashiara | X-Fade can talk about the build-server, for the web etc services I guess there is no problem if you use a hosts -file | 20:14 |
qwazix | yep, the way I get it is they're meant to stay that way | 20:14 |
DocScrutinizer05 | fine! :-) | 20:14 |
qwazix | and only the autobuilders will go on the new machines | 20:15 |
DocScrutinizer05 | qwazix: AIUI nope, they are meant to move to our RX300 once they are ready | 20:15 |
DocScrutinizer05 | mashiara: ^^^ ? | 20:15 |
qwazix | mashiara, please clarify, is my understanding correct or Doc's? | 20:15 |
mashiara | autobuilders will go to the physical machines, other services stay at the VMs | 20:16 |
DocScrutinizer05 | mashiara: yep, but where will those VM live? | 20:16 |
qwazix | we'll have to pay extra for that hosting | 20:16 |
mashiara | those live in our Xen "grid" | 20:16 |
DocScrutinizer05 | mhm | 20:17 |
DocScrutinizer05 | so we actually have to pay for those extra, in addition to the RX300 hosting? | 20:17 |
mashiara | The offer we sent to Nokia and which AFAIK you have seen lists the ongoing hosting costs for all the VMs and colo for the two physical machines. | 20:18 |
DocScrutinizer05 | I was just about to quote Bergie's statement that Nemein could get down mere bare bones coLo-costs to 200EUR/machine*month, but that's a quite meaningless figure when actually most of the VM are not living on those RX300 | 20:19 |
DocScrutinizer05 | sorry, we haven't seen any such offer, that's why we're really speculating all the time. We basically got *no* data | 20:20 |
mashiara | ah... | 20:20 |
DocScrutinizer05 | even worse: no idea what we will have to manage and pay for in a month or two | 20:21 |
qwazix | If it's the disk space that's the limiting factor, couldn't we sponsor the addition of a couple of TB to the RX300's? | 20:21 |
qwazix | so that the costs stay at 400/month? | 20:21 |
mashiara | I will need to do some calculations on what the kind of HW required to run all your VMs would cost | 20:22 |
DocScrutinizer05 | qwazix: please don't start too much with those 400/month, it's bare minimum cost to keep the power on on those two servers | 20:22 |
qwazix | ok, rephrasing: ? so that the costs are limited to the CoLo cost of the two RX300 | 20:23 |
DocScrutinizer05 | anyway, until recently (actually 2 days ago) my impression been that board does those negotiations and hw definition and all, and wouldn't like too much contribution from anybody else. 2 days ago however SD69 "instructed" council to come up with a complete plan for the future infra, starting at hw specs and sw setup, and going til funding basically | 20:25 |
mashiara | There is also the nontrivial amount of work for setting up the HW for hosting VMs etc. We can make you some sort of offer (factoring in the HW Nokia is willing to pay) | 20:26 |
qwazix | my impression had been the same, and thus my reply to Rob, | 20:26 |
DocScrutinizer05 | please let's not start talking offers here, I prefer this getting done in private, I had a lengthy nice chat with Bergie this afternoon, and we need to be careful not to make up any numbers that have no substance | 20:27 |
mashiara | definitely, and calculating one would take more time anyway | 20:28 |
qwazix | In order for nemein to do the math once and for us to be able to have a meaningful summary of the possible costs I think we should first come up with the two extremes | 20:28 |
DocScrutinizer05 | it's pretty visible to everybody what the cash of board looks like right now | 20:28 |
qwazix | on one end keeping everything as is and fixing whatever is broken | 20:28 |
qwazix | and on the other keeping the absolutely necessary | 20:29 |
mashiara | I was kinda wondering why bergie was asking for the price of raw rackspace since it's meaningless for you. | 20:29 |
DocScrutinizer05 | qwazix: i'm already doing exactly this, if you wouldn't speed so much | 20:29 |
*** freemangordon has joined #maemo-meeting | 20:29 | |
qwazix | Ok, I'm sorry, I'll let you finish first | 20:29 |
DocScrutinizer05 | mashiara: yep, that's what I realize now too, when i notice we have a lot of VM not living on those 2 servers | 20:30 |
DocScrutinizer05 | qwazix: as mentioned above I also think that negotiations about "best possible offer" are hardly done in public | 20:31 |
DocScrutinizer05 | that's been the reason why i had that chat with bergie this afternoon | 20:31 |
qwazix | I understand, but I'm not talking financials, just a discussion of what can be ditched | 20:31 |
DocScrutinizer05 | that's about our options | 20:32 |
DocScrutinizer05 | for listing our options we first need a comprehensive summary where we are | 20:32 |
qwazix | :nod: | 20:32 |
mashiara | there are 12 VMs, totaling about 48GB of RAM and 2TB of disk (on SAN, RAIDed etc) allocated. | 20:34 |
DocScrutinizer05 | ok, afaik Nokia pays for the migration work done by Nemein. they pay for hosting in January and February (of the whole infra incl the VM I guess. mashiara?) and they have some small slice for "knowlege transfer" in that budget | 20:34 |
mashiara | Correct. | 20:35 |
DocScrutinizer05 | so far I thought all this would live on the 2 dedicated servers Nokia donates to community | 20:35 |
DocScrutinizer05 | for a ballpark number: a managed vserver costs roundabout 250bucks per month | 20:36 |
DocScrutinizer05 | I now learn we got 12 of them | 20:36 |
DocScrutinizer05 | :x | 20:37 |
mashiara | 11 if you discount the temporary builder VM | 20:37 |
DocScrutinizer05 | brkn: correct about the costs? | 20:38 |
brkn | ack | 20:39 |
DocScrutinizer05 | thanks | 20:39 |
DocScrutinizer05 | so unless Nemein is doing a *massive* discount on this, we definitely need to find a different solution since right now we couldn't afford this setup for a single month | 20:40 |
mashiara | I will need to check what was the hosting cost we quoted to Nokia | 20:40 |
DocScrutinizer05 | I'm strongly favoring to move all VM on the dedicated 2 servers | 20:40 |
Woody14619 | my understanding was that most of those vms were low-use and could be co-hosted on a single box. (Thus the 3 servers, 2 real builders, 1 for the rest as VMs or in a merge point). | 20:41 |
DocScrutinizer05 | Woody14619: ack | 20:41 |
DocScrutinizer05 | same here, when i looked at that pdf document about maemo infra migration | 20:42 |
DocScrutinizer05 | from last year May | 20:42 |
DocScrutinizer05 | dsumming up the VM listed there looked like they were sized and tailored to fit exactly on the dedicated servers | 20:43 |
DocScrutinizer05 | Woody14619: latest info is we get two servers | 20:44 |
DocScrutinizer05 | and I think that should suffice, regarding the somewhat lower load we see these days, compared to when maemo fremantle was "young" | 20:44 |
* Woody14619 nods | 20:45 | |
Woody14619 | I'd bet the dev community would be willing to put up with slightly slower builds if it means keeping more community tools (like garages, wiki, etc) | 20:45 |
DocScrutinizer05 | btw sorry to everybody for me not finding the time to send out a proper mail summing up all those details beforehand. There been little time from this afternoon's chat with bergie until now | 20:46 |
DocScrutinizer05 | Woody14619: exactly. bergie stated he thinks as last extreme resort we could consolidate *all* but autobuilder on one vserver, >>but that wouldn't be any fun anymore for community<< | 20:47 |
DocScrutinizer05 | basically RAM and CPU grunt are just for convenience | 20:48 |
DocScrutinizer05 | stuff will still sort of "work" even on a relatively small hardware - it just gets incredibly slow | 20:48 |
DocScrutinizer05 | a hard limit is storage needed | 20:49 |
DocScrutinizer05 | brkn: correct? | 20:49 |
brkn | depends on what you are intending to store. | 20:49 |
mashiara | there is still time to re-spec the physical boxes btw, but the work for migrating everything *again* needs to be paid for (or done by volunteers). | 20:50 |
DocScrutinizer05 | :nod: | 20:50 |
DocScrutinizer05 | quite unfortunate | 20:50 |
DocScrutinizer05 | though I'm no expert, I'd think moving a VM to a new server and start it there under KVM or XEN shouldn't be *that* hard | 20:51 |
Woody14619 | We host 6 or 8 VMs (one per core) on similar hardware where I work. They're not that bad when you don't break beyond the core. And some of these VMs have far lower priority/traffic CPU wise than anything we do here. | 20:52 |
qwazix | Which of the services have their own VM? | 20:53 |
DocScrutinizer05 | thedead1440: brkn: any comments? | 20:53 |
brkn | i second qwazix, a list of VMs and what they do would be helpful | 20:54 |
qwazix | I'd very like to see a list of VM's ... | 20:54 |
DocScrutinizer05 | sure, definitely | 20:54 |
thedead1440 | ^^^ me too; a proper list needs to be properly published | 20:54 |
brkn | + average load on those VM, else it's just guessing | 20:54 |
Woody14619 | These use and E5-2600 family. Even the low-end processors are 6 core (12 threads). Putting 6 VMs on a box seems reasonable. 12 may be a bit much, but... | 20:55 |
mashiara | currently the VMs are idle since there is no traffick due to DNS changes taking *much* longer than anticipated | 20:56 |
mashiara | https://docs.google.com/document/d/1NNbnYAqAd6PYouPAV2w-13MJ5GFYpapHwS2omyUWGTE/edit | 20:57 |
DocScrutinizer05 | aah yes, the DNS issue - mashiara could you send me, or publish, the IPs, so devels could start using whatever is actually up and running there? | 20:57 |
mashiara | scratchbox actually can also be counted out from the 11 since I guess it's not foundations problem | 20:57 |
DocScrutinizer05 | I'm not sure about scratchbox and who's going to pay for it | 20:58 |
qwazix | what is static.maemo.org? | 20:59 |
mashiara | for serving static files | 20:59 |
DocScrutinizer05 | Woody14619: even if we'd host 20 VM on a dualcore, they would swap like mad but basically still "work" (whatever work means), no? | 20:59 |
mashiara | for the website | 20:59 |
Woody14619 | A couple of those may want a litlte more umph (www/db), but some are quite light-weight and don't see heavy traffic (mail/bugs/test). | 21:01 |
qwazix | mashiara, :nod: | 21:01 |
DocScrutinizer05 | mashiara: the document of May is similar to your URL and (c) Ferenc/Niels. May you allow me publishing it, just for background info? | 21:01 |
qwazix | Unfortunately we can't ditch services easily with that architecture. | 21:02 |
qwazix | Which may be needed later on if our funds diminish | 21:02 |
Woody14619 | I'm betting the 9 or 10 listed could run on a single RX300 rather nicely. | 21:02 |
DocScrutinizer05 | I don't think "ditching services" earns us much | 21:02 |
qwazix | So maybe we should reorganize the structure a little bit so that things are more modular | 21:03 |
DocScrutinizer05 | Woody14619: ack | 21:03 |
Woody14619 | The only one I worry about is www, since right now that's just midguard and friends. Not TMO. | 21:03 |
mashiara | DocScrutinizer05: please send the link to me privately first, I'm not quite sure which document you refer to :) | 21:03 |
DocScrutinizer05 | ok, np | 21:03 |
mashiara | ref TMO reggie was sceptical about running that in VM | 21:04 |
DocScrutinizer05 | let me put it this way: either we try running it *somehow* or we give up already | 21:05 |
* Woody14619 nods | 21:05 | |
qwazix | I'm not sure I agree with that | 21:06 |
DocScrutinizer05 | ok, here's what been originally discussed, it has some more info - though obsolete: http://maemo.cloud-7.de/hidden/maemo.orginfrastructureforHildonFoundation.pdf | 21:06 |
Woody14619 | I'm sure it would much prefer it's own home, but a shared home is better than no home. | 21:06 |
qwazix | I'd prefer to keep tmo, wiki and builder only running than have this behemoth struggle | 21:06 |
Woody14619 | builder, without a unified source (garage) or target (repo) is a nightmare. Ask the harmattan community... | 21:07 |
DocScrutinizer05 | qwazix: we need to find the right balance and possibly even the right sized servers once stuff is running again | 21:07 |
DocScrutinizer05 | qwazix: no use in prematurely shutting down services | 21:07 |
Woody14619 | That's one of the nice parts about VMs, you can scale them rather easily post-fact. | 21:08 |
DocScrutinizer05 | exactly | 21:08 |
Woody14619 | If TMO needs 3 cores, or more space, that can be added dynamicly. | 21:08 |
Woody14619 | And/or other services can be shifted to other physical systems. | 21:09 |
DocScrutinizer05 | and worst thing to happen when resources are limited is everything coming to a grinding halt until admin issues 3 or 4 commands | 21:09 |
* Woody14619 nods | 21:10 | |
qwazix | Woody14619, but the way it is now, with the db server on a seperate vm and the dmz it's hard to say, e.g. we move the wiki to a cheap hosting server elsewhere... | 21:10 |
Woody14619 | No, it's not... in fact it's easier. | 21:10 |
qwazix | or shut one specific thing down because it's broken and we have no maintainer | 21:10 |
DocScrutinizer05 | we for example even could give autobuilder low priority, so it slows down to let other services run at acceptable speed for interactive | 21:10 |
Woody14619 | You can shut off one VM a lot easier than turning off a service on a real system and hope it doesn't have side effects. | 21:11 |
qwazix | Woody14619, yes, if the VM is encapsulated | 21:11 |
qwazix | as it is services are scattered around daisy-chained vm's | 21:12 |
Woody14619 | Most of them are, or at least are grouped somewhat logically. | 21:12 |
DocScrutinizer05 | anyway, I want to avoid wasting mashiara's time with discussions that are not relevant for him | 21:12 |
brkn | Woody14619: factor in administration overhead, even one time overhead for setting up cacti, nagios, munin or sth else | 21:12 |
brkn | to monitor all your vms | 21:12 |
DocScrutinizer05 | mashiara: could we get the VM's image files on those servers? | 21:13 |
Woody14619 | brkn: true, but that's a one time cost. And from the sounds of it, it's something we're going to need to do anyway. | 21:13 |
mashiara | ref IPs: https://docs.google.com/spreadsheet/ccc?key=0AnO-uv1h34ahdC1aUmtXR2xNQk1BbHR2VjFOOHVXUFE this is the document we use to track what DNS changes are requested | 21:13 |
mashiara | ignore the notes, they're basically our internal shorthand notes | 21:14 |
DocScrutinizer05 | mashiara: let me rephrase: do you see any problems when we request to move the VM images files to the server, so we could set up KVM or whatever for them? | 21:14 |
DocScrutinizer05 | mashiara: (IPs) thanks! :-) | 21:15 |
mashiara | There should be no problem of giving you the raw disk images | 21:15 |
DocScrutinizer05 | good | 21:15 |
DocScrutinizer05 | this raises the question: who's savvy enough and willing to set those VM up on those servers, on a voluntary basis? | 21:16 |
Woody14619 | brnk: plus, in reality, we're probably going to have separate teams working on things broken out similarlly to the way the VMs are setup. Admining in that instance will be smaller scope. | 21:17 |
brkn | Woody14619: so each sub-admin will be responsible for the security of his own vm? | 21:17 |
DocScrutinizer05 | Woody14619: that's exactly the plan: get maintainers and sub-admins for each VM, basically | 21:18 |
DocScrutinizer05 | brkn: well, that's something sub-sysops need to coordinate with global sysop | 21:18 |
Woody14619 | brkn: we wil have policies for security on each, but in essence, yes. With one or two over-arching admins to check and/or consult with if it's outside their relm of understanding. | 21:18 |
brkn | i see | 21:19 |
DocScrutinizer05 | Woody14619: exactly | 21:19 |
Woody14619 | We're not looking to replicate the "one man to run them all" architecture. :) | 21:19 |
DocScrutinizer05 | it's just the question who's going to be our master of administration, the almighty savvy super-sysop | 21:20 |
Woody14619 | In 5 to 10 years, we may be down to one or two admins.. but we'll probably be down to 1 box by then too, the size of a salt cube, running on solar power on the global wifi... ;) | 21:21 |
DocScrutinizer05 | since he's the one who's going to set up the VMs for us | 21:21 |
Woody14619 | Not necesseraly so Doc. I'm sure there are more than a few of us familiar with VM setup/admin. | 21:22 |
Woody14619 | .oO(Is this the spirited round of "not it" you were refereing to earlier? ) :) | 21:22 |
DocScrutinizer05 | mashiara: I tried IP for wiki.ameo.org of your list, and ended on bugs.maemo.org - is that correct or did I commit a typo? | 21:24 |
Woody14619 | wiki/bugs are the same IP in the list. | 21:24 |
DocScrutinizer05 | ooh | 21:25 |
DocScrutinizer05 | lol | 21:25 |
Woody14619 | likely only one is the "default" | 21:25 |
DocScrutinizer05 | yeah | 21:25 |
Woody14619 | If you spoof your local dns, or hand-craft the GET with the proper name, it will probably multi-home you to the right place. ;) | 21:25 |
DocScrutinizer05 | etc7hosts must do | 21:26 |
mashiara | you will need to use the names to access the boxes, setup your hosts file for that | 21:26 |
DocScrutinizer05 | :nod: | 21:26 |
DocScrutinizer05 | I'll prepare a "howto" | 21:26 |
DocScrutinizer05 | regarding that offer to Nokia I never seen, could you forward it to council@maemo.org please? | 21:27 |
mashiara | I'll ask bergie to do that. | 21:27 |
DocScrutinizer05 | thanks :-) | 21:28 |
Woody14619 | and/or board@maemo.org as I'm sure the will want to see it as well, if they haven't already. | 21:28 |
DocScrutinizer05 | now, is there anything community/council can do to help Nemein in the migration process? | 21:28 |
DocScrutinizer05 | Woody14619: thanks, I almost forgot | 21:29 |
DocScrutinizer05 | or rather, I assumed they have it and forgot to show to council | 21:29 |
DocScrutinizer05 | dunno why I was assuming that | 21:29 |
mashiara | anything else requiring me ? | 21:32 |
DocScrutinizer05 | mashiara: now, is there anything community/council can do to help Nemein in the migration process? | 21:32 |
DocScrutinizer05 | mashiara: also would you mind coming here next week same time again? | 21:34 |
mashiara | next week might be a problem but I'll try | 21:34 |
mashiara | as for help, mainly there is the question of what to do with the physical HW | 21:34 |
DocScrutinizer05 | and of course, many thanks for joining and sharing info and answering all our questions | 21:34 |
DocScrutinizer05 | mashiara: we'll have discussion about the phy servers and come up with some suggestion, much appreciating your comment on whether it looks semi-reasonable or not | 21:35 |
DocScrutinizer05 | (we == council, community, board) | 21:36 |
DocScrutinizer05 | if that's OK for you | 21:37 |
mashiara | I'm not sure how hung Nokia/board is on the "two physical boxes" thing, basically the budget is 7k EUR excluding VAT | 21:37 |
DocScrutinizer05 | for HW or for complete thing | 21:38 |
mashiara | for HW | 21:38 |
DocScrutinizer05 | I guess if any feasible, we rather sooner than later get rid of them and swap them for good money that can buy us "normal" hosting plus some sysop hours | 21:38 |
qwazix | That would be the best solution | 21:39 |
DocScrutinizer05 | btw that's basically an idea bergie said was reasonable in our chat | 21:39 |
qwazix | We could run for a year on that budget, if not more | 21:40 |
* mashiara wonders if bergie has any idea of re-sale value (or loack thereof) of HW | 21:40 | |
mashiara | unless you can renegotiate the HW to cash for the foundation you will not be able to convert the machines to 7k of cash. | 21:41 |
DocScrutinizer05 | he didn't mention any resale values, he just said it would of course have been the better option if Nokia donated money for hosting, but they decided to donate HW | 21:41 |
qwazix | What bothers me is why Nokia decided to go that way in the first place | 21:41 |
DocScrutinizer05 | and "but you're probably free to do whatever you like, after february - or even sooner" | 21:42 |
mashiara | sure, it's your HW after february | 21:42 |
qwazix | Do we know specs of that boxes? | 21:43 |
DocScrutinizer05 | so how are we going to tackle the issue of stopping purchase of that HW, and rather get hosting for same amount of money? | 21:43 |
qwazix | ram, hdd, cpu? | 21:43 |
mashiara | qwazix the specs are to be defined, original suggestin was specced for build, so RAM and CPU heavy | 21:44 |
DocScrutinizer05 | qwazix: see link I posted | 21:44 |
mashiara | so you can still re-spec within the budget | 21:44 |
DocScrutinizer05 | http://maemo.cloud-7.de/hidden/maemo.orginfrastructureforHildonFoundation.pdf | 21:44 |
qwazix | I see that the build service isn't too heavily used the past months | 21:44 |
qwazix | Maybe we're overdesigning? | 21:45 |
Woody14619 | My thoughts exactly. | 21:46 |
M4rtinK | well, since November the autobuilder interface was broken | 21:46 |
DocScrutinizer05 | we most likely are | 21:46 |
M4rtinK | so that limited builder usage quite a bit | 21:46 |
DocScrutinizer05 | qwazix: where from you get that info? | 21:46 |
qwazix | Or we could chip in a couple of hundred bucks and get 4tb of diskspace for one of the two servers | 21:46 |
DocScrutinizer05 | sure | 21:46 |
qwazix | DocScrutinizer05 nothing scientific, but I was watching the "last build" thingie on maemo.org | 21:47 |
DocScrutinizer05 | but that assumes we're keeping the servers | 21:47 |
DocScrutinizer05 | aaah | 21:47 |
qwazix | even when the builder was working sometimes the latest thing built was over a day ago | 21:47 |
DocScrutinizer05 | mashiara: who's doing the server puchase in Nemein? | 21:48 |
mashiara | me | 21:48 |
DocScrutinizer05 | ok, so please don't hurry | 21:48 |
mashiara | :) | 21:48 |
qwazix | that means it was sitting there idle all the other time, no? | 21:48 |
DocScrutinizer05 | qwazix: that's kinda correct but autobuilder considered "stopped2 no, for good | 21:49 |
DocScrutinizer05 | qwazix: that's kinda correct but autobuilder considered "stopped" now, for good | 21:49 |
DocScrutinizer05 | aiui | 21:49 |
qwazix | yep, and maybe I was wrong thinking that it worked because I was using scp, while others couldn't upload through the web-interface | 21:50 |
qwazix | so disregard my comment above | 21:50 |
DocScrutinizer05 | mashiara: am I correct with autobuilder? | 21:50 |
kerio | 7k of servers is too much, really | 21:51 |
qwazix | then again 64gb of memory sounds too much for our builder. It's not that we have daily builds of the whole system or something | 21:51 |
mashiara | I'm sorry, I don't know, X-Fade does. | 21:51 |
DocScrutinizer05 | kerio: no, it's way too little, for 700k the sysop is built-in | 21:51 |
DocScrutinizer05 | mashiara: :nod: | 21:51 |
kerio | like, a dwarf sysop living inside the server? | 21:51 |
DocScrutinizer05 | yup | 21:51 |
kerio | neat! | 21:52 |
DocScrutinizer05 | anyway, please don't disturb meeting | 21:52 |
qwazix | It's just people building individual packages, delays IMHO are acceptable | 21:52 |
M4rtinK | also scp doesn't work for some (including me) | 21:52 |
DocScrutinizer05 | anymore comments/questions to Nemein / mashiara ? | 21:52 |
qwazix | (I believe that most of us just use the builder for publishing and not for testing while developing, no?) | 21:53 |
DocScrutinizer05 | qwazix: that's the idea | 21:53 |
kerio | how much does a *hosted* server cost at nemein? | 21:53 |
Woody14619 | Some do use it for that, but it's rare.. and again, I don't see issues with that process being slower if it means other stuff works. :) | 21:53 |
DocScrutinizer05 | kerio: you can ask this in a private mail to nemein | 21:54 |
qwazix | re comments/questions, nothing more from me | 21:54 |
Woody14619 | kerio: scroll back.. already covered. | 21:54 |
Woody14619 | I'm good | 21:54 |
Woody14619 | I do have a Council topic, when we get back to that. ;) | 21:54 |
mashiara | kerio: all prices quoted for maemo are "special price, only for you" | 21:54 |
DocScrutinizer05 | mashiara: many thanks for participating | 21:54 |
kerio | i see | 21:55 |
mashiara | we like these guys but cannot afford to sponsor all this... | 21:55 |
DocScrutinizer05 | kerio: that's why I said we don't want to discuss prices in public here | 21:55 |
qwazix | mashiara, this is understandable | 21:56 |
mashiara | anyway, I'm going now. email me if you need something. | 21:56 |
qwazix | bye! | 21:56 |
DocScrutinizer05 | will do, thanks again mashiara! :-) | 21:56 |
mashiara | np, bye. | 21:56 |
*** mashiara has left #maemo-meeting | 21:56 | |
DocScrutinizer05 | hope to see you next week | 21:57 |
DocScrutinizer05 | oops gone | 21:57 |
DocScrutinizer05 | ok | 21:57 |
DocScrutinizer05 | Woody14619: ? | 21:57 |
Woody14619 | So... :) One topic to throw in on, just to keep it alive: Status of last 2 wandering CA awards? | 21:57 |
DocScrutinizer05 | hmm, that's qwazix, no? | 21:57 |
qwazix | Woody14619, last thing was before christmas | 21:58 |
qwazix | let me find the latest email | 21:58 |
DocScrutinizer05 | bbiam | 21:58 |
qwazix | our contact said it might be impossible to find the lost N9's but that he would try to source N950's instead | 21:59 |
qwazix | I will ping kojacker to see if there was anymore correspondence between them | 22:00 |
Woody14619 | Ok... Interesting. :) Curious that they don't have some tracking system, since most of the others got theirs with clear tracking info... | 22:01 |
qwazix | sorry, it was just after christmas... | 22:01 |
qwazix | so taking into consideration holidays and such, not many days have passed since | 22:02 |
DocScrutinizer05 | re | 22:02 |
qwazix | Some just found the device at home on the other hand | 22:02 |
Woody14619 | Understood. Just wanted to make sure it didn't get lost in the holiday shuffle. | 22:02 |
qwazix | I'll talk to kojacker anyway to get latest info | 22:03 |
Woody14619 | Well, to my knowledge there were only 2 in the missing category. :) Good enough! | 22:03 |
qwazix | It was 2/100, not really bad | 22:04 |
qwazix | (considering all devices in the program) | 22:04 |
qwazix | Do we have anything else? | 22:05 |
thedead1440 | can member of public raise an issue wrt TMO moderation? | 22:05 |
DocScrutinizer05 | sure | 22:06 |
Woody14619 | Usually? :) | 22:07 |
DocScrutinizer05 | this is a free and public meeting | 22:07 |
DocScrutinizer05 | thedead1440: shoot! | 22:07 |
thedead1440 | its a small thing I wanted to put out for debate: I feel that once all this is settled for TMO moderation it should be looked into that Council/Board members are not appointed moderators/super moderators during their terms | 22:08 |
thedead1440 | this would allow anybody in TMO to debate with them without having the proverbial sword hanging over them | 22:08 |
thedead1440 | that since that person is a mod he could do a ban etc | 22:08 |
Woody14619 | I think the only person that has any moderator privs that's on the board/council at the moment is TexRat. No? And he's had that position for well over 3 years. | 22:09 |
DocScrutinizer05 | i'd think that's self-evident | 22:09 |
thedead1440 | timsamoff? | 22:09 |
thedead1440 | so previously it was 2/3rd of Board | 22:09 |
thedead1440 | now 1/3rd of Board | 22:09 |
DocScrutinizer05 | ok, good point | 22:10 |
Woody14619 | I'm unaware if Tim has and TMO privs? | 22:10 |
DocScrutinizer05 | didn't really realize | 22:10 |
thedead1440 | he's a super moderator to boot | 22:10 |
DocScrutinizer05 | he has | 22:10 |
thedead1440 | not just moderator like Texrat | 22:10 |
DocScrutinizer05 | afaik at least | 22:10 |
thedead1440 | he has i checked it before raising this ;) | 22:10 |
Pali | DocScrutinizer05, I looked to that dns file on google docs and there is missing new entry for A record: downloads.maemo.org | 22:10 |
Woody14619 | In both cases though, they were moderators well before they were Board/Council. | 22:11 |
DocScrutinizer05 | Pali: you're the greatest, like always! :-) Thanks, will let Eero know | 22:11 |
Woody14619 | And both were Council in that time frame as well... | 22:11 |
thedead1440 | so my point here is during their terms they should hold only 1 position not both | 22:11 |
Pali | DocScrutinizer05, now it has same A record as www.maemo.org | 22:11 |
kerio | hm, did nokia tell us if they're going to keep hosting tablets-dev? | 22:12 |
Pali | maemo.org. 82 IN A 80.248.164.250 | 22:12 |
Pali | downloads.maemo.org. 43107 IN A 80.248.164.250 | 22:12 |
Woody14619 | kerio: Nokia never hosted tables-dev. That was all Reggie. | 22:12 |
DocScrutinizer05 | see freemangordon's mail correspondence with SD69 | 22:12 |
kerio | really? :o | 22:12 |
kerio | Woody14619: isn't that like... the ofifcial firmware release site? | 22:13 |
Pali | Woody14619, are you sure that nokia did not hosted tablets-dev.nokia.com ?? | 22:13 |
DocScrutinizer05 | Woody14619: I think you talk about internettablettalk | 22:13 |
Woody14619 | My bad... I misread. Yes, tablets-dev is Nokia. | 22:14 |
Woody14619 | thedead1440: While I understand your concern, I'm not sure there's a good solution for this. | 22:14 |
DocScrutinizer05 | we clearly need to push this issue thru board | 22:14 |
thedead1440 | My proposed solution is that during Board/Council term said BoD or Councillor should not hold any moderator position on TMO | 22:15 |
Woody14619 | Realistically, moderators were chosen based on their temperment and ability to keep a level head. | 22:15 |
DocScrutinizer05 | reasonable | 22:15 |
Woody14619 | Not really... | 22:15 |
DocScrutinizer05 | though normal moderators can still get overruled by supermod and admin | 22:15 |
thedead1440 | well since we are on it I would like to request to know what exactly timsamoff has been super moderating the past few months | 22:16 |
Pali | DocScrutinizer05, is nokia going to give us full maemo.org domain? | 22:16 |
Woody14619 | With the dwindling resources we're facing, limiting how many hats a person can wear will artificially limit us. | 22:16 |
DocScrutinizer05 | Pali: they are supposed to do that | 22:16 |
thedead1440 | he said in his nomination that he was returning to TMO; well that means before that he had a power but wasn't available to use it | 22:16 |
thedead1440 | he returned to lead the NFP via the Board and still retains his power | 22:16 |
thedead1440 | once everything is done and dusted he will leave and keep his position again? | 22:17 |
thedead1440 | So is this a for-life title? | 22:17 |
Woody14619 | thedead1440: I'm not really clear what's going on here. IS there a specific issue at hand? | 22:17 |
thedead1440 | Nope seriously no issue at all | 22:17 |
thedead1440 | I just was remembering Estel's case | 22:17 |
DocScrutinizer05 | Woody14619: still I'd think a councilor or board member should refrain from any moderation in case of conflicts with somebody about the moderator's other positions. Much like I finally tried on IRC wrt estel | 22:17 |
Woody14619 | thedead1440: This "title" was given by Reggie, who was the sole operator and proprieter of TMO. | 22:17 |
thedead1440 | when he said that Texrat was behind it | 22:17 |
thedead1440 | just because Texrat was being elected to Board | 22:18 |
thedead1440 | so I feel such a conflict should not manifest itself again | 22:18 |
Woody14619 | Estel is frankly a lying sack of shit... Sorry to be blunt... | 22:18 |
DocScrutinizer05 | as a general rule no moderator ever should morate conflicts he's peronsally involved, no matter what other hats he got | 22:18 |
Pali | DocScrutinizer05 and is TMO going to be migrated too? | 22:18 |
DocScrutinizer05 | Pali: yes | 22:18 |
Woody14619 | His initial ban was by another moderator... | 22:18 |
Pali | because I do not see it in that doc... | 22:18 |
thedead1440 | yes I know it was by chemist | 22:19 |
Woody14619 | And his month long ban was done by Reggie himesel. | 22:19 |
thedead1440 | Again yes I know that | 22:19 |
Woody14619 | For breaking clearly set forum rules. | 22:19 |
thedead1440 | What I mean is for an outsider looking in or someone looking to raise a point in the future it doesn't seem very clear cut on what they can debate on safely | 22:19 |
Pali | DocScrutinizer05, I tested login on new migrated maemo wiki and seems it working :-) | 22:20 |
Woody14619 | Sorry, but I think cripling ourselves (losing moderators) to keep up some type of "outside" appearance for the sake of someone who did more damage to this community while he was here than ANY moderator has ever done, is kind of stupid. | 22:20 |
DocScrutinizer05 | Pali: that's yet another issue we need to sort out, since originally the assumption been we can move it to a VM on the RX300 | 22:21 |
thedead1440 | Well I wanted to raise a point for general debate | 22:21 |
DocScrutinizer05 | Pali: \o/ | 22:21 |
thedead1440 | In your case of losing moderators like I said I don't see much timsamoff moderated while he was away i.e. not part of Board or nomination process | 22:21 |
Woody14619 | I agree, that if there is some kind of direct conflict... (eg moderation is needed on a discussion ABOUT or involving Council/Board, or existing moderators) there should be a policy to give that task to a non-involved moderator. | 22:22 |
DocScrutinizer05 | thedead1440: it's a good point indeed, but I'm afraid we can't solve it any better than by stating what's best practice | 22:22 |
Woody14619 | But AFAIK that's already something in place. | 22:22 |
thedead1440 | Yup hope that is indeed the case then | 22:22 |
Woody14619 | It may be a good idea, once Board takes up operation of TMO, to review moderators and/or write up something more specific policy wise... | 22:23 |
DocScrutinizer05 | we will make it go in place when it's not and the problem comes up | 22:23 |
DocScrutinizer05 | just like more reasonable co-chanops instructed me about how to behave re estel on #maemo | 22:24 |
DocScrutinizer05 | :-) | 22:24 |
DocScrutinizer05 | I'm pretty much a friend of common sense in that regard | 22:24 |
Woody14619 | Having that in loosly written wording would always be good... (Perhapse a good task, for Council to draft?) | 22:24 |
DocScrutinizer05 | sure, I think MT will put it on the action-list when doing the minutes | 22:25 |
Woody14619 | But really, making a strict policy of one person shall not hold 2 hats, is going to do nothing but cripple the community as we shrink. | 22:25 |
DocScrutinizer05 | yup | 22:25 |
DocScrutinizer05 | anyway, I didn't open the meeting. So i'm not going to close it now, before I leave ;-) | 22:26 |
Woody14619 | We did something along those lines already for Board.... Discouraging holding two seats at once... With one exception that was handled by the community in voting. :) | 22:26 |
DocScrutinizer05 | unless somebody has further points to discuss with me | 22:26 |
Woody14619 | (And that exception I feel has been more of a help than a detractor... to the community anyway.) | 22:26 |
Woody14619 | I'm good. :) | 22:27 |
DocScrutinizer05 | just one more comment - a wise man in another channel posted: >>or 1 month the VMs as they currently are; are allowed to run to find out their load etc since Nokia are paying for it so its 0 cost to the community. once you have the figures end of month 1 then get alternative solutions including selling away of the 2 servers<< | 22:29 |
kerio | we won't get back all the money for the servers though | 22:30 |
DocScrutinizer05 | yep, that's why I asked Eero to not hurry with buying them | 22:30 |
DocScrutinizer05 | mentalist: action-pont. approach board to see if we can re-negotiate with Nokia | 22:31 |
DocScrutinizer05 | kerio: mind you, for now the VM are NOT running on "our2 servers, and they are not even planned to do so | 22:32 |
kerio | why the double move though? | 22:32 |
kerio | :s | 22:32 |
Woody14619 | I think the issue at hand is that Nokia doesn't own the servers they're running now. They're leasing them. | 22:32 |
kerio | no, nvm, ignore me | 22:33 |
Woody14619 | That would mean when the lease is up, they go away. The concept here being they buy hardware to put this all on, and we pay service costs. | 22:33 |
DocScrutinizer05 | sure, that's been the big-business idea nad solution | 22:33 |
DocScrutinizer05 | they not even bothered about cost of CoLo etc | 22:34 |
Woody14619 | I think we'll find that even one of these systems will handle most of the load for the VMs, with some space to spare. | 22:34 |
DocScrutinizer05 | I'm deeply convinced somebody inside Nokia expected us to break in tears of joy | 22:34 |
DocScrutinizer05 | Woody14619: ack | 22:35 |
Woody14619 | And splitting the second into Builder and TMO VMs is probably our best solution... | 22:35 |
DocScrutinizer05 | again full ack | 22:35 |
DocScrutinizer05 | with builder low prio, so it cranks up speed when low load on tmo | 22:36 |
Woody14619 | I don't think we're going to get much more (or a conversion from server to $) from Nokia. | 22:36 |
DocScrutinizer05 | I have my doubts as well | 22:36 |
DocScrutinizer05 | anyway I hope we see somewhat clearer now about our options | 22:37 |
Woody14619 | But even with what we have... I think we can pull it all off. We just need to setup a reasonable donnor setup (small donations, $3 to $10 Euro a month), and with enough people, that foots the bill and starts a small kitty for hardware replacement if needed down the road. | 22:37 |
DocScrutinizer05 | we need someone driving this technically | 22:37 |
* Woody14619 nods | 22:38 | |
* DocScrutinizer05 lurks at brkn | 22:38 | |
Woody14619 | Frankly, I'm really happy with where this is moving to. I was not expecting this to work as smoothly as it has. | 22:38 |
*** ZogG_laptop has quit IRC | 22:38 | |
*** ZogG_lap1op has joined #maemo-meeting | 22:39 | |
DocScrutinizer05 | until just yesterday, I was feeling final doom pending from board dropping stuff on council's feet | 22:39 |
Woody14619 | Lots of thanks go to everyone involved... | 22:39 |
DocScrutinizer05 | indeed, from me as well | 22:40 |
DocScrutinizer05 | 160min meeting, but worth every second. I think we got stuff moving a bit | 22:40 |
Woody14619 | :) I think that's not quite the case... If we keep moving, even at the current pace, we'll be in place in time for a bumpy but working cut over. (The bumpy part being now, when things are down and read-only at times) | 22:41 |
Woody14619 | (not quite was referencing your doom, not the productivity of the meeting) | 22:41 |
DocScrutinizer05 | aah :-) | 22:42 |
Woody14619 | K, something is beeping at me... got to go smack something with a stick. :) See you all later. | 22:42 |
DocScrutinizer05 | cya, and thanks | 22:42 |
* DocScrutinizer05 also leaving | 22:43 | |
*** brkn has left #maemo-meeting | 22:43 | |
*** thedead1440 has left #maemo-meeting | 22:43 | |
Pali | DocScrutinizer05, in that doc also missing A record for "maintenance.maemo.org" | 22:46 |
DocScrutinizer05 | isn't maintenance.maemo.org more like an internal thing? | 22:47 |
Pali | it is website with old style | 22:48 |
Pali | DocScrutinizer05, see also bug: https://bugs.maemo.org/show_bug.cgi?id=12581 | 22:48 |
povbot | Bug 12581: Not possible to browse history of Developer forum | 22:48 |
DocScrutinizer05 | hmmm | 22:48 |
DocScrutinizer05 | frankly I have no idea what's supposed to be there | 22:49 |
Pali | you can see that website with old style has buttons for history | 22:50 |
Pali | but new not | 22:50 |
Pali | and old is on maintenance.maemo.org | 22:50 |
DocScrutinizer05 | yeah, but what *is* "history"? | 22:50 |
Pali | older posts | 22:52 |
Pali | on website with old style you have buttons "1" "2" "3" "4" "5" "6" "next" ">>" | 22:52 |
Pali | see screen shots on that bug report | 22:53 |
*** Pali has quit IRC | 22:57 | |
DocScrutinizer05 | well, i still have them | 22:57 |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!