Time |
Nick |
Message |
16:02 |
dogi |
#topic agenda |
16:03 |
|
* last meeting actions review |
16:03 |
|
- housetree/munin fixed - http://sunjammer.sugarlabs.org[…]ugarlabs.org.html |
16:03 |
|
- more ram for pootleVM - http://sunjammer.sugarlabs.org[…]s.org-memory.html |
16:03 |
|
- turn lightwaveVM to master dns |
16:03 |
|
- new oleVM - http://ole.sugarlabs.org is not http://library.ole.org/ |
16:03 |
|
- update on confcall with RIT folks |
16:03 |
|
- update on shipping of 9 machines to MA/DC |
16:03 |
dfarning |
+1 for etherpad |
16:03 |
dogi |
* short munin overview for treehouse/housetree servers |
16:03 |
|
http://sunjammer.sugarlabs.org[…]ugarlabs.org.html |
16:03 |
|
http://sunjammer.sugarlabs.org[…]ugarlabs.org.html |
16:03 |
|
- what was the gap in the munin log of treehouse around 23:xx yesterday? |
16:03 |
|
- does pootle still needs more ram? |
16:04 |
|
* new meetingVM dogi |
16:04 |
|
* new monitoringVM (munin/openQRM) dogi |
16:04 |
|
* new ubuntu-sugarteam buildmachine dfarning |
16:04 |
|
* new trac (on housetree?) bernie |
16:04 |
|
* redhat/fedora offers VMspace ... discuss |
16:04 |
|
dfarning, thx :) |
16:04 |
|
#link http://etherpad.com/fRDL07G6d3 |
16:04 |
|
<- meeting agenda link on etherpad |
16:05 |
|
so |
16:05 |
|
now |
16:05 |
|
welcome to the weekly treehouse meeting |
16:05 |
|
hi all |
16:05 |
decause |
afternoon gents |
16:05 |
dogi |
hi decause |
16:05 |
decause |
and ladies for that matter |
16:06 |
|
waves to dogi |
16:06 |
dogi |
http://me.etin.gs/treehouse/tr[…]0100126_1601.html |
16:06 |
dfarning |
do I know decause or CDeslandes |
16:06 |
dogi |
<- backlog for you, decause |
16:06 |
decause |
decause = RemyD, Resident Hacktivist at RIT |
16:06 |
|
dogi: thanks |
16:06 |
dogi |
hi silbe |
16:06 |
|
http://me.etin.gs/treehouse/tr[…]0100126_1601.html |
16:07 |
silbe |
thx! |
16:07 |
decause |
CDeslandes: CSH sysadmin type, main contact for Sugar/CSH collab |
16:07 |
dogi |
only lfarone is missing ... |
16:07 |
dfarning |
awesome welcome |
16:07 |
decause |
catches up on backlog... |
16:08 |
dogi |
is there something to add? |
16:08 |
silbe |
CSH was RIT? |
16:08 |
dogi |
think we will make RIT/ASLO point bigger |
16:08 |
decause |
Computer Science House is a student Organization at RIT, that runs their own projects and campaigns at RIT |
16:09 |
silbe |
kthx |
16:09 |
decause |
they have an entire floor of a dormatory, with some pretty impressive computing resources and talent |
16:09 |
dogi |
maybe we do a short intruduction ... |
16:09 |
silbe |
decause, CDeslandes: Welcome, BTW! |
16:09 |
dfarning |
+1 |
16:09 |
decause |
they've been strong for 35 years now, and will be the new loving home for the 3U's that you are sending |
16:09 |
dogi |
is Stefan Unterhauser, Systems Admin for SL |
16:10 |
|
dfarning, ? |
16:10 |
dfarning |
Dfarning is a goofy guy trying to start a service and support business for Sugar |
16:10 |
dogi |
:P |
16:10 |
silbe |
is Sascha Silbe, SL sysadmin as well (mainly buildbot.sugarlabs.org, sugar-jhbuild) |
16:10 |
decause |
_bernie: hi there |
16:11 |
_bernie |
hello |
16:11 |
|
decause: ciao! |
16:11 |
tonyf |
hello |
16:11 |
dfarning |
hey _bernie made a meeting |
16:12 |
silbe |
hi tonyf! backlog: http://me.etin.gs/treehouse/tr[…]0100126_1601.html ; we're currently doing introduction round |
16:12 |
dogi |
thx silbe |
16:12 |
dfarning |
tonyf could you introduce yourself |
16:12 |
_bernie |
dfarning: haha :) |
16:12 |
tonyf |
hi tony forster, melbourne australia |
16:12 |
silbe |
dfarning: maybe we should schedule the other meetings equally late? ;) |
16:13 |
|
hmm, that name rings a bell, just not sure which one exactly... |
16:13 |
tonyf |
most meetings are 3am local time :( |
16:13 |
decause |
lolz |
16:13 |
dogi |
so think we are ready for next pont |
16:13 |
|
#topic last meeting actions review |
16:13 |
dfarning |
tonyf: sorry didn't remember your nick |
16:14 |
dogi |
housetree/munin fixed - http://sunjammer.sugarlabs.org[…]ugarlabs.org.html |
16:14 |
|
thx _bernie :) |
16:14 |
|
more ram for pootleVM - http://sunjammer.sugarlabs.org[…]s.org-memory.html |
16:14 |
|
now 2GB |
16:15 |
silbe |
I think someone complained about it getting OOM today |
16:15 |
|
was that before or after your change? |
16:15 |
dogi |
there it look we need more or more stabilty ... lets discuss that late |
16:15 |
|
rr |
16:15 |
|
later |
16:15 |
silbe |
ok |
16:15 |
dogi |
new oleVM - http://ole.sugarlabs.org is not http://library.ole.org/ |
16:16 |
_bernie |
dogi: cjl said today that pootle was running out of memory |
16:16 |
dogi |
thx chrowe |
16:16 |
_bernie |
dogi: was it before or after you added 1GB? |
16:16 |
silbe |
_bernie: check your backlog ;) |
16:16 |
dogi |
and RIT/CSH folks are here |
16:16 |
|
that was the prelast point |
16:17 |
|
thx lfarone |
16:17 |
|
ups |
16:17 |
|
lfaraone |
16:17 |
decause |
nod |
16:17 |
dogi |
#topic short munin overview for treehouse/housetree servers |
16:18 |
|
http://sunjammer.sugarlabs.org[…].org.html#Virtual machines |
16:19 |
silbe |
shouldn't we discuss a.sl.o first so we don't waste the time of our guests? |
16:19 |
dogi |
there was a munin log break |
16:19 |
|
and we have some highs |
16:19 |
|
#action dogi look into break |
16:20 |
|
#action _bernie lock into highs? |
16:20 |
|
ok _bernie |
16:20 |
|
hi lfaraone |
16:20 |
|
:) |
16:21 |
|
nearly the same is on http://sunjammer.sugarlabs.org[…].org.html#Virtual machines |
16:21 |
silbe |
dogi: about the munin blackouts: |
16:21 |
|
17:40 < bernie> cure: btw, do you know why sunjammer was rebooted yesterday? |
16:21 |
|
18:11 < nvjacobo> bernie: because the colo facility had problems yesterday |
16:21 |
lfaraone |
dogi: hi, I'm going to be afk almost immediately :( |
16:21 |
dogi |
thx silbe |
16:21 |
lfaraone |
is gone. |
16:21 |
dogi |
lfaraone, next? |
16:22 |
decause |
If I may gents, I know that we're pre-last point on the agenda, but I'd like to talk RIT/CSH Logistics sooner than later if we could :P |
16:22 |
dogi |
decause, you are the next point |
16:22 |
decause |
:P |
16:22 |
dogi |
#topic lfaraone RIT ASLO |
16:22 |
dfarning |
dogi: lfaraone left I think |
16:23 |
decause |
lfaraone: still here? |
16:23 |
dogi |
so who will start dfarning ? |
16:23 |
decause |
well, we talked last night briefly |
16:23 |
|
I'll scare up the text |
16:23 |
_bernie |
<18:20> * lfaraone is gone. |
16:23 |
|
decause: well, I guess you can talk to me |
16:23 |
dogi |
ok _bernie |
16:24 |
decause |
_bernie: heres what we discussed so far: |
16:24 |
CDeslandes |
1) Where should he send the hardware/who will be responsible for receiving it? |
16:24 |
|
2) What is the uplink/network connection like? |
16:24 |
|
3) these computers should hopefully have more than one ethernet port, would it be possible to have them to, in addition to the primary internet link, be on their own private net or daisy-chained together? that way we can avoid putting excess load when synchronizing the machines. |
16:25 |
decause |
these are the 3 issues that lfaraone brought to me and CSH last night |
16:26 |
CDeslandes |
For point one, we would like the hardware to be sent to the CSH address, sent to the chairman, with attention to Christopher Deslandes |
16:26 |
dogi |
+1 |
16:27 |
CDeslandes |
For point 2 we currently have a 1 GB uplink to the RIT backbone, unthrottled. This provides access to the Internet and Internet2 |
16:27 |
_bernie |
CDeslandes: I think point 1 is no problem |
16:27 |
CDeslandes |
We can talk to the IT dept. on campus and see if we can get another pipe run if necessary. |
16:27 |
|
_bernie: ok |
16:27 |
dfarning |
CDeslandes: how is the reliability? |
16:27 |
silbe |
CDeslandes: Internet2? |
16:27 |
_bernie |
CDeslandes: Internet2 == ipv6? (sorry if I'm being ignorant) |
16:28 |
dfarning |
dfarning: capable of handling a tier one service? |
16:28 |
CDeslandes |
Academic only Internet, Universities, Hospitals, etc, also much faster than standard internet |
16:28 |
dfarning |
what about summers and vacations? |
16:28 |
CDeslandes |
Here's a link for more info: http://www.internet2.edu/ |
16:29 |
silbe |
ah, I see, thx. |
16:29 |
CDeslandes |
The server room has access during the summer, normally one sysadmin stays in the area |
16:29 |
dfarning |
+1 |
16:29 |
dogi |
+1 |
16:29 |
_bernie |
silbe, dogi: what's going on with disk space on sunjammer? http://sunjammer.sugarlabs.org[…]rlabs.org-df.html |
16:30 |
CDeslandes |
For vacations, normally one of us is in the area, the only time where it would be completely unattended would be during Christmas break in which the entire campus is locked up |
16:31 |
dfarning |
We will grow enough by next christmas that the school will give you guys and exception:) |
16:31 |
_bernie |
this is also strange: http://sunjammer.sugarlabs.org[…]ysql_threads.html |
16:31 |
CDeslandes |
Currently we do not have IPv6 set up, but we do have accounts with Hurricane Electric for IPv6 tunnels, and it is a project that will most likely come back to surface soon. |
16:31 |
dogi |
_bernie, aslo? |
16:31 |
CDeslandes |
_bernie: :) |
16:31 |
_bernie |
dogi: not sure... |
16:31 |
silbe |
CDeslandes: about the uplink, how much traffic would you tolerate? a.sl.o keeps attracting an increasing number of users, so even though large files are offloaded to mirrors, the traffic might be significant (and growing). |
16:32 |
CDeslandes |
I think we can take a decent amount of traffic, plus we have room to expand, if necessary |
16:32 |
dogi |
decause, CDeslandes that brings me to the next point ... diskspace |
16:32 |
CDeslandes |
We will just have to fight the IT dept. on campus for it |
16:32 |
silbe |
CDeslandes: ok, so our machines saturating the interface wouldn't be an issue? |
16:33 |
decause |
dogi: we're working on putting up a 11 Terabyte SAN now |
16:33 |
|
it won't be up for another 2 weeks or so |
16:33 |
|
for now, diskspace is what we need most |
16:33 |
|
we've got plenty of cycles and bandwidth |
16:34 |
dogi |
thx decause |
16:34 |
decause |
this CSH cloud project we are working on will eventually have the capacity to maybe take on some of the load |
16:34 |
|
but for now, until that cloud is up, we'll stick to the baremetal |
16:35 |
silbe |
decause: sounds good so far |
16:35 |
dogi |
perfect :) |
16:35 |
silbe |
decause: thx for the offer, BTW! |
16:36 |
dfarning |
decause: are you guys aware of the design of a.sl.o are there people interested in keeping it going |
16:36 |
decause |
this is the main thing you guys want to migrate, right? |
16:36 |
|
activities.sugarlabs.org? |
16:36 |
dfarning |
+1 |
16:37 |
dogi |
would like to help |
16:37 |
CDeslandes |
dfarning: We are constantly bringing in people who are interested in Systems Administration |
16:38 |
decause |
we can set up some kind of official relationship (internship/co-op/independent study) to make this happen as well |
16:38 |
dfarning |
awesome, it is pretty cool |
16:38 |
CDeslandes |
dfarning: We normally have more people that want to get involved than what we know to do with. |
16:38 |
dogi |
+1 |
16:38 |
silbe |
decause: yes, it's the service that's in most need of scaling |
16:38 |
decause |
silbe: understood. How much of a draw is it currently? (Disk, bandwidth, etc...?) |
16:39 |
silbe |
because it's used by the actual users, not just volunteers and developers |
16:39 |
dogi |
+1 silbe |
16:39 |
silbe |
decause: it's hosted on sunjammer, so http://sunjammer.sugarlabs.org[…]ugarlabs.org.html might give you a hint |
16:40 |
|
there are also other services on that host, including shell service |
16:40 |
|
the disk space needed for a.sl.o is rather small in comparison |
16:41 |
dogi |
#action look more into http://sunjammer.sugarlabs.org[…]rlabs.org-df.html |
16:41 |
_bernie |
dogi: I'm on it :) |
16:41 |
silbe |
the 80GB drives currently in the WMF servers should be sufficient AFAIK, but local backup storage would be nice to have |
16:41 |
_bernie |
silbe: indeed |
16:42 |
dogi |
_bernie, :) me too |
16:42 |
dfarning |
decause: because it is a pretty php application is it pretty much cpu bound |
16:43 |
dogi |
I think we will should move on ... |
16:43 |
decause |
CDeslandes finish perusing the sunjammer stats |
16:43 |
dogi |
so the only thing which is missing is to send them |
16:43 |
_bernie |
decause, CDeslandes: how much rack space can we hope for? we have these 3 boxes (1U each) plus a fourth one available from RIT itself |
16:44 |
dogi |
good point bernie |
16:44 |
_bernie |
decause, CDeslandes: all of these are used boxes, probably not very reliable |
16:45 |
dogi |
thinks they would be at least good proxies ... |
16:45 |
silbe |
oh right, power switches and potentially serial consoles would be nice to have |
16:45 |
_bernie |
I guess we'll discuss the details of the software strategy when we're settled with the hardware. |
16:46 |
CDeslandes |
We do have a network accessible serial console |
16:47 |
silbe |
decause: what about IPv4 addresses? these seem to start actually running out... |
16:47 |
CDeslandes |
And we can spare 3U to 4U of space, but that is about it, for now |
16:47 |
_bernie |
CDeslandes: serial console would be great, thanks |
16:47 |
silbe |
CDeslandes: great! |
16:47 |
CDeslandes |
CSH currently has 2 external Class C blocks |
16:47 |
dogi |
+1 silbe |
16:47 |
_bernie |
CDeslandes: ok. if these boxes turn out to be too crappy, we'll purchase a new 2U box to replace them |
16:48 |
CDeslandes |
And we are probably using about 50% of that |
16:48 |
_bernie |
CDeslandes: for $3000 we can get an 8 core machine with redundant fans etc |
16:48 |
dogi |
+1 _bernie |
16:48 |
CDeslandes |
_bernie: :D |
16:49 |
silbe |
CDeslandes: ok, I was starting to get worried - our university has a full Class B and the dormotories are even behind NAT |
16:49 |
|
*dormitories |
16:49 |
_bernie |
silbe: this may be due to close-mindedness... I know people in italy who use NAT as some sort of security measure ;-) |
16:49 |
decause |
I've got to prepare for my OLPC Honors Seminar Students shortly, so I've got about 5-10 mins left |
16:50 |
dfarning |
I much perfer the idea of a house full of geeks working on a custer. Me ducks |
16:50 |
CDeslandes |
We also have some dual 2.8 Ghz 1U machines that could be used as a temporary measure if the boxes that are sent decide to die. |
16:50 |
silbe |
_bernie: that's the real reason, not the official one ;) |
16:50 |
CDeslandes |
CSH is behind a transparent firewall |
16:51 |
_bernie |
decause: I guess we have a deal. We'll ask lfaraone to go on with the shipment, ok? |
16:51 |
silbe |
decause: ok, so anything else _you_ would like to know? |
16:51 |
decause |
_bernie: sounds good by me. CDeslandes ? |
16:51 |
dogi |
+1 _bernie |
16:51 |
dfarning |
+1 |
16:51 |
CDeslandes |
_bernie: sound good to me as well |
16:51 |
dogi |
decause, can I visit CSH and RIT? |
16:51 |
decause |
asap |
16:51 |
|
we'll give you a proper tour |
16:52 |
dogi |
:) |
16:52 |
decause |
btw, stallman is going to be here in february, so that *may* be an interesting time to take a holiday... ;) |
16:52 |
silbe |
dfarning: one doesn't rule out the other - it's just going to be a cluster of fat machines instead of a slim ones :) |
16:52 |
_bernie |
decause: tell him hello from me |
16:52 |
dogi |
#action lfaraone machines to RIT |
16:53 |
decause |
_bernie: will do |
16:53 |
dogi |
ok? _bernie? lfaraone ? |
16:53 |
decause |
we're going to be making a really big deal of this once things are officially up |
16:53 |
_bernie |
dogi: ok |
16:53 |
dogi |
decause, :P |
16:54 |
silbe |
decause: you're quite welcome to - we get more than our fair share out of it IMO |
16:54 |
decause |
I'm hanging out in #openinnovation on Freenode, and that is where me and lfaraone handled business yesterday |
16:54 |
_bernie |
decause: do you think we could really find skilled and interested students to help run the cluster? |
16:54 |
silbe |
disclaimer: i'm not a member of the marketing team ;) |
16:54 |
_bernie |
is very worried for reliability |
16:54 |
dfarning |
decause: my goal is to turn CSH in to an http://osuosl.org/ for the east coast! |
16:54 |
decause |
_bernie: every student needs a co-op/internship to graduate |
16:54 |
|
;) |
16:54 |
_bernie |
dfarning: that would be cool |
16:55 |
silbe |
decause: lol |
16:55 |
_bernie |
decause: we'll abuse them as much as possible then |
16:55 |
decause |
its more symbiotic than abuse |
16:55 |
dogi |
lol decause |
16:55 |
CDeslandes |
_bernie: The servers that we currently manage have close to 99% uptime |
16:55 |
decause |
mutually beneficial is how I'd like to think of it :P |
16:56 |
|
and Sugarlabs is one of the highest profile opensource projects in the movement right now |
16:56 |
|
IMHO atleast |
16:56 |
silbe |
decause, CDeslandes: sounds like a great deal to both sides then :) |
16:56 |
decause |
excellent |
16:56 |
|
then I'll be in channel, and looking forward to the next actions |
16:56 |
_bernie |
CDeslandes: but our servers are probably going to be crappy, and it looks like we're going to depend on all 4 of them being up in order to erogate the service |
16:56 |
decause |
_bernie: We'll find a way _bernie, no worries. |
16:57 |
dogi |
thx decause |
16:57 |
|
and thx CDeslandes |
16:57 |
silbe |
decause: what's that channel about? There's no topic set |
16:57 |
CDeslandes |
_bernie: we run zenoss to ensure that our systems are up and running |
16:58 |
|
_bernie: It e-mails us when things go down, and we all have e-mail on our phones |
16:58 |
_bernie |
CDeslandes: great, that would help a lot |
16:58 |
decause |
anything else before I depart gents? |
16:58 |
dogi |
http://www.zenoss.com/ |
16:58 |
|
http://community.zenoss.org/index.jspa |
16:59 |
_bernie |
decause: I think we're good, thanks! |
16:59 |
decause |
_bernie: ok then |
16:59 |
_bernie |
dogi: forgot to say #link? |
16:59 |
dfarning |
decause: sounds good |
16:59 |
dogi |
ups |
16:59 |
silbe |
decause, CDeslandes: have a nice evening (I guess ;) )! |
17:00 |
dogi |
next point |
17:00 |
CDeslandes |
silbe: Thanks, you too |
17:00 |
decause |
dfarning: _bernie: dogi: keep me in the loop. CDeslandes is my go-to guy on CSH, so anything you need to ask me you can ask him tool. |
17:00 |
dogi |
#topic pootle |
17:00 |
silbe |
thx! |
17:00 |
dogi |
ok decause |
17:01 |
|
unmad|away, ping |
17:01 |
|
think he is already asleep |
17:01 |
|
but wanted to ask bernie anyway |
17:01 |
_bernie |
dogi: about what? |
17:01 |
dogi |
do u think pootle need still more ram |
17:01 |
|
??? |
17:02 |
_bernie |
dogi: I think 2GB ought to be enough |
17:02 |
dogi |
http://sunjammer.sugarlabs.org[…]s.org-memory.html |
17:03 |
_bernie |
dogi: memcached seems to be way bigger than needed |
17:03 |
dogi |
hmm |
17:04 |
_bernie |
dogi: it looks like yesterday it ate all memory, probably a leak. |
17:04 |
|
dogi: giving it more memory would just make the problem worse |
17:04 |
dogi |
ok |
17:05 |
_bernie |
unmad|away: when you're back on the keyboard, let us know what you think |
17:05 |
|
next topic? |
17:05 |
dogi |
plans for pootleVM is to move it to housetree |
17:05 |
|
but for that I want stability ... |
17:06 |
|
#topic new trac |
17:06 |
|
#topic new tracVM |
17:06 |
|
_bernie, that is something I wrote on this meeting because of you |
17:07 |
|
do we want to have a devel - trac ? |
17:07 |
|
what does other think? silbe? |
17:08 |
silbe |
dogi: would be nice to have |
17:08 |
|
oh, that reminds me of something, _bernie: admin access to git.sl.o :) |
17:08 |
dogi |
silbe, :) |
17:09 |
silbe |
dogi: we want to move Trac proper as well, BTW. |
17:09 |
dogi |
ok |
17:10 |
silbe |
whether VM or sunjammer doesn't matter much too me, as long as me fiddling with the web server to set up XML-RPC or similar is fine :) |
17:10 |
dogi |
who needs access to this VM silbe |
17:11 |
_bernie |
silbe: oh ys |
17:11 |
dogi |
_bernie, will you create this VM |
17:11 |
_bernie |
dogi, silbe: I'd like someone to take over maintaining trac |
17:11 |
|
I think mchua said she would |
17:12 |
dogi |
is not a trac expert |
17:12 |
_bernie |
the vm may or may not be useful... I don't really care much about it :-) |
17:12 |
silbe |
_bernie: I might try setting it up to run with sth. different from Apache (if it turns out to be faster), but I guess Apache could play proxy if necessary |
17:12 |
|
_bernie: what OS is sunjammer again? |
17:13 |
_bernie |
silbe: ubuntu |
17:13 |
silbe |
ok, at least not Fedora... |
17:13 |
_bernie |
silbe: jaunty, but it will be upgraded to karmic soon or later |
17:13 |
|
silbe: the problem of trac is trac itself. it sucks with or without a vm, with or without a proxy... it just sucks. |
17:14 |
dogi |
#action tracVM with access for mchua silbe bernie dogi |
17:14 |
|
lol _bernie |
17:14 |
silbe |
_bernie: sure, but maybe we can try sucking through a shorter and/or bigger pipe? ;) |
17:14 |
_bernie |
dogi: no, I'm not at all interested in the VM for trac because then I'd have to maintain BOTH trac and its VM :-) |
17:15 |
dogi |
hmm |
17:15 |
silbe |
it's ok, I'll maintain the VM :) |
17:15 |
dogi |
+1 silbe |
17:15 |
_bernie |
silbe: then we just move it on sunjammer. it would take, like 30 minutes of work. |
17:15 |
silbe |
U-break-untu should have unattended-upgrades from Debian, that makes it easier :) |
17:15 |
|
_bernie: ? |
17:16 |
_bernie |
silbe: ubuntu also have unattended upgrades, no? |
17:16 |
dogi |
:) yes and yes |
17:16 |
silbe |
_bernie: they probably have just imported the Debian tool in which case it should actually work ;) |
17:17 |
|
(sorry for making Ubuntu look bad - I've just had a lot of bad experience with Ubuntu distro updates, no matter what tool got used) |
17:17 |
_bernie |
silbe: it says "This script is the backend for the APT::Periodic::Unattended-Upgrade option." |
17:18 |
silbe |
_bernie: then it's the Debian one, good. |
17:18 |
_bernie |
silbe: ubuntu is not as solid as debian, I agree. I think it's a deliberate choice to deliver more recent software. |
17:19 |
silbe |
_bernie: certainly |
17:19 |
dogi |
#topic new ubuntu-sugarteam buildslave |
17:19 |
silbe |
dfarning: what do you need? |
17:20 |
dogi |
thats something I would like to maintain |
17:20 |
|
bu there i need help |
17:20 |
|
ups but |
17:20 |
dfarning |
Just a very small vm running lucid |
17:21 |
_bernie |
dfarning: you're giving us a good excuse to upgrade our template |
17:21 |
silbe |
dfarning: what are you (currently) going to do on it? Does reliability matter? Lots of bandwidth? |
17:21 |
dogi |
#action new lucidVM dogi |
17:21 |
_bernie |
dfarning: would karmic upgraded to lucid through apt work for you, or do you want a clean installation from the lucid cd? |
17:21 |
dogi |
lol bernie but right now lucid is on alpa 2 |
17:21 |
|
alpha |
17:22 |
dfarning |
silbe: I build them on my desktop, reliability is not an issue. |
17:22 |
_bernie |
dfarning: (I don't think it matters much as long as we're careful about what packages we add) |
17:22 |
silbe |
dfarning: do you need root access? would you mind other users on the same VM? |
17:23 |
dfarning |
silbe: one a day the box with need to pull update and push a new iso |
17:23 |
_bernie |
dogi: yes, I'm proposing to create a new template based on ours (not throw away ours :-) |
17:23 |
dogi |
sure _bernie that was the plan :) |
17:23 |
dfarning |
silbe: others are fine, but the build pocess does require root. |
17:24 |
dogi |
rewrite template recipe for lucid alpha 2 |
17:24 |
silbe |
dfarning: DVD or CD image (just trying to get an estimate of the upstream usage) |
17:24 |
|
? |
17:24 |
dogi |
dfarning, good to know |
17:24 |
_bernie |
dogi: you're volunteering to this? -- great! :) |
17:24 |
dogi |
lol yes |
17:25 |
|
but like I said it could be I need help ... _bernie |
17:25 |
dfarning |
cd, but I think I can set up zsync to make the iso syncs very small |
17:25 |
silbe |
dogi: I can help you, no problem. |
17:25 |
dogi |
and silbe |
17:26 |
silbe |
_bernie: that spec would match a new VM (with buildbot) on bender as well. what do you think? |
17:26 |
_bernie |
dogi: I'll help |
17:26 |
dogi |
think with that the meeting was long enough |
17:26 |
|
that remembers me _bernie |
17:26 |
_bernie |
dfarning: zsync? hmm interesting |
17:26 |
dogi |
http://etherpad.com/Sm3YTAwvLb |
17:27 |
dfarning |
_bernie: http://linux.die.net/man/1/zsync it works very well for incrementing iso |
17:27 |
_bernie |
dfarning: /me reads |
17:27 |
dogi |
which name should we have for this VM |
17:27 |
|
buildbot5? |
17:27 |
_bernie |
dogi: template-lucid? |
17:27 |
silbe |
dogi: buildslave-ubuntu-lucid-64bit |
17:27 |
_bernie |
dogi: ah, yes... what silbe said |
17:28 |
dogi |
ok bernie for template |
17:28 |
_bernie |
it's a whopping long name for my taste, but let's follow the established convention :) |
17:28 |
silbe |
(hooray for bash configured to autocomplete host names <g>) |
17:28 |
dogi |
and ok silbe for first clone |
17:29 |
silbe |
_bernie: so bender or treehouse? |
17:29 |
dogi |
#topic others |
17:29 |
_bernie |
silbe: you remind me of a programmer friend who was using variable names of 30-40 chars and then recommending everyone to buy microsoft visual studio with the visual assist extension. |
17:29 |
|
:-) |
17:29 |
dogi |
silbe, treehouse |
17:29 |
silbe |
_bernie: except I try to cause a buffer overflow in MS products :-P |
17:29 |
_bernie |
dfarning: if you have no needs for bandwidth and can access through ipv6, I would recommend using bender for a build box |
17:29 |
dogi |
moved last 3 points to next week |
17:30 |
_bernie |
dfarning: oh, wait... you have to transfer isos around? then treehouse :) |
17:30 |
silbe |
ah yes, there was a catch we bender. I always assume everybody has IPv6 by now ;) |
17:30 |
|
*with |
17:31 |
dogi |
by the way tonyf would u like to add something? |
17:31 |
_bernie |
dfarning: I mean, housetree |
17:31 |
silbe |
_bernie: what's the difference anyway? |
17:31 |
_bernie |
silbe: treehouse is being used for production services so I'd rather not run builds concurrently on it |
17:32 |
silbe |
ah, OK. |
17:32 |
_bernie |
silbe: housetree is relatively idle at this time. right dogi? |
17:32 |
silbe |
that's why i thought of bender instead :) |
17:32 |
dfarning |
IPV6 give my local isp a head ache, I live in rural wisconsin:) |
17:32 |
dogi |
jupp _bernie |
17:32 |
silbe |
dfarning: I'm using a tunnel for IPv6 upstream as well |
17:32 |
|
I think only the netherlands have good native IPv6 coverage |
17:33 |
dogi |
but let me first create the buildmachine then we can always move the VM later |
17:33 |
_bernie |
dfarning: try ping 192.168.90.2 |
17:33 |
|
dogi: +1 |
17:33 |
dogi |
is that thing between sunjammer and treehouse fixed? |
17:34 |
silbe |
oh, BTW: _bernie: do you mind me setting up squid on bender some time? |
17:34 |
dfarning |
_bernie: it just hangs |
17:34 |
dogi |
to remind the bottleneck on housetree is there are only 5 IPs there |
17:34 |
_bernie |
silbe: I don't mind, but why do you need it? |
17:35 |
silbe |
sunjammer and treehouse/housetree are at the same physical locations? (sorry, I can't seem to remember the locations) |
17:35 |
_bernie |
dfarning: ok then your isp really doesn't have a tun6to4 gateway |
17:35 |
|
dfarning: if you still want ipv6 at home, you could use a free sixxs.net account or a free he.net account |
17:36 |
dogi |
dogi sunjammer:~$ ssh treehouse.crosslink |
17:36 |
|
ssh: connect to host treehouse.crosslink port 22: No route to host |
17:36 |
silbe |
_bernie: just to reduce traffic from the build slaves. I'd like to set up nightly Debian package builds and want to make sure they're not going to pull the same packages over and over :) |
17:36 |
_bernie |
dfarning: another possibility: you can log in on sunjammer and then ssh to any ipv6 host from there |
17:36 |
dogi |
silbe, yes treehouse is next to sunjammer |
17:36 |
|
and housetree next to solarsail |
17:36 |
_bernie |
silbe: ok good plan |
17:36 |
silbe |
dogi: good to know, thx! |
17:36 |
_bernie |
silbe: develer also has squid running on trinity btw |
17:37 |
dfarning |
_bernie: yes, that is what I generally do:) |
17:37 |
dogi |
silbe, :) |
17:37 |
_bernie |
silbe: are you planning to use transparent proxying over ipv6? :-) |
17:37 |
silbe |
_bernie: ah good, will try to set it as parent then |
17:37 |
|
_bernie: no thx, i hate transparent proxies :) |
17:37 |
_bernie |
silbe: then just use trinity as a proxy without installing anything on sunjammer |
17:37 |
|
silbe: if it doesn't love you, I'll change acls |
17:38 |
silbe |
_bernie: but maybe a tinydns+dnscache instance so we can use local names ;) |
17:38 |
_bernie |
silbe: s/sunjammer/bender/ |
17:38 |
silbe |
_bernie: ok, will do. |
17:39 |
_bernie |
silbe: I'm not *against* running squid on bender, I just think it may be less work to reuse trinity's |
17:39 |
silbe |
_bernie: does squid on trinity have refresh-ims set? |
17:39 |
dfarning |
anything else for me? other wise I am off to scheme up more partnerships:) |
17:39 |
silbe |
(reminds me that I need to investigate why it isn't working as expected on flatty :) ) |
17:39 |
_bernie |
silbe: to answer your old question: sunjammer and treehouse are at the same location (GNAPS), housetree is at the Media Lab. |
17:39 |
silbe |
dfarning: see in in IRC tomorrow? |
17:40 |
|
*you in |
17:40 |
_bernie |
silbe: what is refresh-ims? |
17:40 |
silbe |
_bernie: oh, so housetree is physically separate? i thought they were twins... |
17:41 |
_bernie |
silbe: they are twins, not siamese twins |
17:41 |
dfarning |
silbe: Yes, I have a phone call with alaxander. Then you and I can follow up on IRC |
17:41 |
dogi |
ping chrowe |
17:41 |
silbe |
_bernie: tells squid to ensure a fresh entry if the client uses If-Modified-SiC[C[C[C[C[C[C[Cnce |
17:41 |
|
If-Modified-Since that should have read |
17:42 |
|
I'm having issues with squid giving false "unmodified"s to apt... |
17:42 |
chrowe |
dogi: pong |
17:43 |
dogi |
#topic oleVM |
17:43 |
dfarning |
thanks guys /cu |
17:43 |
silbe |
dfarning: ok, great! you can warn alexander that I'm physically UTC+1 and logically even later (but will stand up early as needed, of course) ;) |
17:43 |
chrowe |
dogi: yes |
17:43 |
silbe |
ok, too late |
17:43 |
_bernie |
silbe: it's the default iirc |
17:44 |
dogi |
http://ole.sugarlabs.org/ |
17:44 |
|
nice video u have there |
17:44 |
silbe |
_bernie: i had to add it explicitly |
17:44 |
|
dogi: what are the remaining topics? |
17:45 |
|
(trying to judge whether to stay or go to bed) |
17:45 |
dogi |
this is the last |
17:45 |
silbe |
ah, ok |
17:46 |
dogi |
thx chrowe |
17:46 |
|
#endmeeting |