torrent migration, reversed
by Seth Vidal
Hi folks,
I attempted to migrate the torrent to the new server today. That didn't
pan out
First, the new torrent version didn't work well under any significant
load. It worked okay in the tests I had run, but not when hit by all of
our users.
I worked around this by playing with the version a bit.
It was working for a while.
Then it died b/c, apparently, the machine it is hosted on died.
I've migrated the hostname back to the old machine and brought the old
torrent back online.
I've also requested the hardware be replaced in the new machine.
As soon as we can get a stable machine I'll attempt the migration,
again.
Thanks,
-sv
--
I only speak for me.
16 years, 2 months
Meeting Log - 2008-04-03
by Ricky Zhou
15:02 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Who's here?
15:03 * dgilmore is here
15:03 * daMaestro is here
15:03 * f13
15:04 -!- fugolini [n=francesc(a)host86-200-dynamic.7-87-r.retail.telecomitalia.it] has quit Remote closed the connection
15:05 * iWolf lurks
15:06 < mmcgrath> pretty low turnout
15:06 * mmcgrath pings some people
15:07 < mdomsch> hi
15:07 < dgilmore> mmcgrath: can we add our CA discussion to the meeting
15:07 < mmcgrath> dgilmore: sure
15:07 < abadger1999> hEY GUYS
15:07 < mmcgrath> lets start with the tickets
15:07 < dgilmore> gday abadger1999
15:07 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Tickets
15:07 < mmcgrath> .tiny https://fedorahosted.org/fedora-infrastructure/query?status=new&status=as...
15:07 < zodbot> mmcgrath: http://tinyurl.com/2hyyz6
15:07 < mmcgrath> Audio streaming for conference calls
15:07 < mmcgrath> .ticket 395
15:07 < zodbot> mmcgrath: #395 (Audio Streaming of Fedora Board Conference Calls) - Fedora Infrastructure - Trac - https://fedorahosted.org/projects/fedora-infrastructure/ticket/395
15:08 < mmcgrath> .any jcollie
15:08 < zodbot> mmcgrath: jcollie was last seen in #fedora-meeting 6 days, 7 hours, 20 minutes, and 27 seconds ago: *** jcollie has quit IRC ("Ex-Chat")
15:08 < mmcgrath> jcollie's not been around in a bit
15:08 < mmcgrath> so we'll skip that one for now.
15:08 < mmcgrath> .ticket 398
15:08 < zodbot> mmcgrath: #398 (elfutils `monotone' (mtn) error) - Fedora Infrastructure - Trac - https://fedorahosted.org/projects/fedora-infrastructure/ticket/398
15:08 * lmacken is here
15:08 < mmcgrath> someone in #fedora-devel mentioned there's a fix for this but I've not seen it yet
15:08 < mmcgrath> .any roland
15:08 < zodbot> mmcgrath: I have not seen roland.
15:09 < mmcgrath> roland was recently added to the sysadmin-hosted group. He may be able to better assist in this
15:09 < dgilmore> .any rmcgrath
15:09 < zodbot> dgilmore: I have not seen rmcgrath.
15:09 < mmcgrath> we'll move on for th
15:09 < mmcgrath> .ticket 446
15:09 < zodbot> mmcgrath: #446 (Possibility to add external links on spins page) - Fedora Infrastructure - Trac - https://fedorahosted.org/projects/fedora-infrastructure/ticket/446
15:09 < mmcgrath> skvidal: ping
15:09 -!- sebastian^ [n=sebastia(a)Ya1e5.y.pppool.de] has joined #fedora-meeting
15:10 < sebastian^> hi ho all
15:10 < mmcgrath> I never really heard anything more on this.
15:10 -!- buggbot [n=supybot(a)landfill.bugzilla.org] has quit Read error: 104 (Connection reset by peer)
15:10 < mmcgrath> IIRC we talked about adding a link to the spins.fp.o page that links to the wiki
15:10 -!- geroldka [n=Gerold@fedora/geroldka] has joined #fedora-meeting
15:10 < mmcgrath> is that what others recall?
15:10 < dgilmore> i think thats the best option
15:10 < mmcgrath> we can always go back and look at the meeting minutes.
15:11 < dgilmore> thats what we talked about
15:11 < mmcgrath> <nod>
15:11 < mmcgrath> dgilmore: would you mind getting that wiki page setup?
15:11 < dgilmore> mmcgrath: can do
15:11 < mdomsch> is spins.fp.o that hard to edit now?
15:11 < mmcgrath> solid.
15:11 -!- buggbot [n=supybot(a)landfill.bugzilla.org] has joined #fedora-meeting
15:11 < mmcgrath> mdomsch: not really hard, no.
15:12 < mmcgrath> alrighty, anyone have anything else on that?
15:12 < mmcgrath> next ticket
15:12 < mmcgrath> .ticket 485
15:12 < zodbot> mmcgrath: #485 (Schedule Koji DB vacuum) - Fedora Infrastructure - Trac - https://fedorahosted.org/projects/fedora-infrastructure/ticket/485
15:12 < mmcgrath> abadger1999: ^^
15:13 < mdomsch> mmcgrath, then I just am curious why do it in the wiki instead of just letting the right folks edit the page directly
15:13 < mdomsch> oh well moving on
15:13 < mmcgrath> mdomsch: well, I think there were a couple of reasons. At least one of which is technical.
15:13 < abadger1999> So we just need to do a complete vacuumof koji.
15:13 < mmcgrath> the other was more about hosting and stuff. How do we know the difference between an "official" spin and a non official one? Do we care?
15:13 < mmcgrath> abadger1999: how long will it take?
15:14 < abadger1999> It will probably make kojia bit slow.
15:14 < abadger1999> Unknown how long it will take but
15:14 < mdomsch> mmcgrath, official ones are blessed by the board, hosting is determined by rel-eng based on space available determined by infra
15:14 < abadger1999> Lasttime we attemptedthis, did it go for longerthan 24 hours?
15:14 < mmcgrath> abadger1999: and it does not require a lock right?
15:14 < abadger1999> Correct.
15:15 < mmcgrath> mdomsch: but how do the users figure it out if they're all listed in the same place?
15:15 < abadger1999> But it could slow things down.
15:15 < mmcgrath> abadger1999: is that something we could just start on Sat?
15:15 < mmcgrath> and hope its done by monday?
15:15 < abadger1999> We can definitely startit.
15:15 < mmcgrath> f13: ping
15:15 < mmcgrath> you guys have any major rebuilds planned for this weekend?
15:15 < mdomsch> listed != hosted :-)
15:16 < dgilmore> mmcgrath: i moved the sparc postgres to 8.3 this week
15:16 < mmcgrath> dgilmore: how'd that go?
15:16 < dgilmore> mmcgrath: ill see how it works since autovacuum is supposed to work
15:16 < f13> mmcgrath: this weekend? Not that I know of, except that it's the last weekend before the Final Freeze for Fedora 9
15:17 < mmcgrath> dgilmore: excellent.
15:17 < dgilmore> mmcgrath: it transitioned smoothly, you have to do a dump restore to change versions
15:17 < mmcgrath> f13: we're thinking about running a vacuum on the db so we stop getting notices about it.
15:17 < mmcgrath> and because we're running out of transactions
15:17 < dgilmore> mmcgrath: i think this weekend is fine
15:17 < f13> hrm.
15:18 < f13> it
15:18 < mmcgrath> dgilmore: <nod>
15:18 < f13> it'll just make things a bit slower right, not break anything?
15:18 < dgilmore> f13: yes
15:18 < mmcgrath> f13: correct.
15:18 < abadger1999> f13: Correct.
15:18 < f13> ok...
15:18 < f13> I reserve the right to be bitchy if something falls over and holds up Fedora 9 builds (:
15:19 * abadger1999 notes that the warnings start when we hit half way so we could go for a bit longer if we have to.
15:19 < mmcgrath> abadger1999: does this weekend work for you or would you rather me do it?
15:19 < mmcgrath> abadger1999: and how the hell did we run out of FAS tranasctions so quick :) ?
15:19 -!- spoleeba [n=one@fedora/Jef] has quit "Leaving"
15:20 < abadger1999> I can do it this weekend
15:20 < abadger1999> FAS is kinda worrisome but at leastit's a small db.
15:20 < mmcgrath> yeah
15:20 < mmcgrath> abadger1999: solid, well let me know if you need anything
15:21 < mbonnet> abadger1999: remember to disable the other vacuums while it's going
15:21 < abadger1999> Right. mbonnet -- only for that db, though, correct?
15:21 < mbonnet> abadger1999: correct
15:21 < abadger1999> Cool. I think we're all set on that then.
15:22 < mbonnet> abadger1999: also, don't bother with the analyze, since that can be done on a table-by-table basis later
15:22 < abadger1999> <nod>
15:22 -!- Southern_Gentlem [n=notfred@unaffiliated/southerngentlem/x-2894754] has joined #fedora-meeting
15:22 < mmcgrath> cool
15:22 < mmcgrath> alrighty, anyone have anything else on that?
15:23 < dgilmore> not I
15:23 < mmcgrath> cool, then we're done with the tickets
15:23 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Certificates
15:23 -!- spoleeba [n=one@fedora/Jef] has joined #fedora-meeting
15:23 < mmcgrath> dgilmore: want to take the floor on that one?
15:23 < dgilmore> mmcgrath: :) yep
15:24 < dgilmore> so looking at all the options for setting up and managing our certificate infrastructure
15:24 < dgilmore> all of them have definete downsides
15:25 < dgilmore> I really dont want to write our own
15:25 < dgilmore> i think its too much effort for too little gain
15:25 < mdomsch> dgilmore, agreed
15:25 -!- kital [n=Joerg_Si@fedora/kital] has joined #fedora-meeting
15:25 < dgilmore> i'm thinking that we should try make dogfood work
15:26 < dgilmore> i have spoken with the devs of it and they commited to helping us
15:26 < skvidal> mmcgrath: pong
15:26 < dgilmore> dog tag
15:26 < skvidal> sorry, was in another desktop
15:26 < mmcgrath> skvidal: I forget what I was going to ask :)
15:26 < skvidal> torrent?
15:26 < mmcgrath> dgilmore: is it in Fedora already?
15:26 < mmcgrath> skvidal: maybe
15:26 < dgilmore> mmcgrath: not yet
15:27 < dgilmore> mmcgrath: however none of the options is
15:27 < dgilmore> one of them requires jboss
15:27 < dgilmore> so it will be a huge task to get in
15:27 < dgilmore> dogtag has some pieces in fedora
15:27 < dgilmore> it uses fedora-ds for storage
15:27 -!- fab_away [n=bellet(a)bellet.info] has quit No route to host
15:28 < dgilmore> the option using a relational db uses jboss
15:28 < dgilmore> i think it was openca
15:28 < dgilmore> one of the other options uses openldap
15:28 < mmcgrath> So whats the scope of this project? We're looking to give and revoke certificates to our users.
15:28 < mmcgrath> will this also sign our VPN and builder certs?
15:28 < dgilmore> yes
15:29 < mmcgrath> will this have anything to do with package signing?
15:29 < dgilmore> nothing to do with packag signing
15:29 < dgilmore> so we will need to setup the ca
15:29 < mmcgrath> Solid, so what do you need to get a proof of concept setup?
15:30 < dgilmore> test that when we revoke certs apps honour the revokation
15:30 < dgilmore> probably will need 2 or 3 vm's
15:30 < mmcgrath> k, the german box telia1 is up
15:30 < dgilmore> setup as many apps as we can
15:30 < mmcgrath> .ping telia1.fedoraproject.org
15:30 < zodbot> pong
15:31 < mmcgrath> We're going to stick one app and one proxy server out there, the rest can be used for test.
15:31 < mmcgrath> we've got a /28 for it.
15:31 < dgilmore> :) nice
15:31 < mmcgrath> if you want to virt-inst some guests over there, have at it.
15:31 < dgilmore> ok will do
15:32 < dgilmore> I'm going to need some help getting things setup and running
15:32 < dgilmore> and testing
15:32 -!- geroldka [n=Gerold@fedora/geroldka] has quit "Leaving"
15:32 < dgilmore> in the end all users will need new certs. as will all builders and vpn endpoints
15:32 < mmcgrath> I'm a little confused on how this will tie in to FAS2 (if it even will)
15:32 < mmcgrath> Like, where will users go to actually get their cert?
15:33 < dgilmore> they will go to fas2
15:33 < mmcgrath> perhaps we'll have a better view after we get it all setup
15:33 < dgilmore> but fas2 will request it from the ca
15:33 < iWolf> dgilmore: i might have time to help with some setup if you need it.
15:33 < dgilmore> iWolf: thanks
15:34 < mmcgrath> solid.
15:34 < mmcgrath> dgilmore: you have anything else on that front?
15:34 < dgilmore> so thats how i think we should procede
15:34 < dgilmore> mmcgrath: i think thats it
15:35 < mmcgrath> solid
15:35 < mmcgrath> dgilmore: sounds good, let me know if you need anything
15:35 < mmcgrath> Movin on
15:35 < dgilmore> mmcgrath: willd o
15:35 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Torrent
15:35 < mmcgrath> skvidal: mind giving a quick overview of what we did, why we did it and how much space we've got left?
15:36 < skvidal> yes
15:36 < skvidal> okay
15:36 < skvidal> 1. eventually got torrent1.fp.o to stay up and online
15:36 < skvidal> 2. allocated 400gb of the 425gb available to /srv of torrent
15:36 < skvidal> 3. installed bittorrent 5.2.0 - had it fall over every 20 minutes or so,
15:37 < skvidal> 4. cursed a lot, attempted to fix
15:37 < skvidal> 5. found it was just not reliable
15:37 < skvidal> 6. rebuilt bt 4.4.0 and installed it
15:37 < skvidal> 7. had it fall over , while not a lot, a fair bit,
15:37 < skvidal> 8. figured out that 4.4.0 and modern python-twisted don't get along
15:37 < skvidal> 9. disable twisted
15:37 < skvidal> 10. restart bttrack
15:37 < skvidal> 11. profit
15:38 < mmcgrath> all in all though we're in good shape now?
15:38 < skvidal> 12. removed zod and bordeaux (among other crap like fedora7-rc2) from torrents
15:38 < mmcgrath> eww
15:38 < skvidal> currently there are 203G remaining
15:38 < skvidal> and it appears to be running correctly
15:38 < skvidal> time and f9 will tell if it is truly reliable
15:38 < mmcgrath> <nod>
15:38 < mmcgrath> skvidal: thanks!
15:39 < skvidal> but atm it is more reliable than it was before
15:39 < mmcgrath> excellent.
15:39 < mmcgrath> anything else?
15:39 < skvidal> I'm updating the SOP, now
15:39 < skvidal> oh and
15:39 < skvidal> if anyone wants a project that will make them instantly excellent
15:39 < skvidal> fork bt 4.4.0 and maintain it/make it sane
15:40 < skvidal> hell, even if you only made it a server (bttrack/bttseed) you'd be a hit
15:40 < skvidal> that's all
15:40 < mmcgrath> :)
15:40 < mmcgrath> skvidal isn't kidding btw :)
15:40 < mmcgrath> alrighty. Movin on
15:40 < dgilmore> skvidal: what is the deal with bt 5.x?
15:41 < skvidal> dgilmore: it would either run out of files (even when I increased it to ludicrous amounts) or it would stop sending the /announce page, at all.
15:41 < skvidal> and then 5.2.0 is the dead end
15:41 < skvidal> bittorrent.com stopped the open source releases
15:41 < dgilmore> :( didnt they change the license?
15:41 < skvidal> oh they're chock full of evil
15:42 < mmcgrath> pure evil
15:42 < mmcgrath> ok, movin on
15:42 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Me gone tomorrow and saturday.
15:43 < mmcgrath> I've got a conference where I'm representin' Red Hat and Fedora.
15:43 < mmcgrath> I'll be back on Sunday.
15:43 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Open Floor
15:43 < mdomsch> enjoy
15:43 < mmcgrath> Ok, so thats really all we had planned for this meeting.
15:43 < mmcgrath> The wiki migration is moving along but there's still TONS to be done for those of you python coders interested in helping, the migration script could always use work.
15:44 < mmcgrath> I'm also working on getting the docbooko stuff all figured out.
15:44 < mmcgrath> ricky: you around by chance?
15:45 < mmcgrath> guess not.
15:45 < mmcgrath> we're still blocking a bit on FAS2 and OpenID but I'm pretty confident that'll be in good shape.
15:45 < mmcgrath> Alrighty, anyone have anything else they'd like to discuss? If not I'll close the meeting.
15:45 < dgilmore> mmcgrath: making fas a openid provider?
15:46 * dgilmore is done
15:46 < mmcgrath> dgilmore: yeah
15:46 < poelcat> mmcgrath: what is the targeted golive date for the new wiki?
15:46 < iWolf> mmcgrath: where is the migration script for the wiki?
15:46 < mmcgrath> poelcat: don't have one yet but it will be after F9 ships. Hopefully by a week or two.
15:46 < poelcat> okay
15:46 < mmcgrath> iWolf: git.fedorahosted.org/git/fedora-infrastructure.git/scripts/
15:47 < iWolf> mmcgrath: thanks!
15:47 < mmcgrath> alrighty, if no one has anything else we'll close the meeting in 30
15:47 < mmcgrath> 15
15:47 < mdomsch> mmcgrath, when is freeze for f9?
15:47 < mmcgrath> mdomsch: ahh, good question.
15:48 * mmcgrath looks.
15:48 * mdomsch sneaks in under the wire
15:48 < mmcgrath> mdomsch: well last time we did a 2 week freeze IIRC.
15:48 -!- gkrpan [n=greg(a)c-71-56-252-37.hsd1.co.comcast.net] has joined #fedora-meeting
15:49 < mmcgrath> f13: did you want a longer change freeze this time around?
15:49 < mdomsch> yes, and that seemed to work pretty well
15:49 < mmcgrath> we could do a partial change freeze on the 8th followed by a full freeze on the 15th?
15:49 * mmcgrath doesn't know what a partial change freeze is but we could come up with something.
15:49 < mmcgrath> otherwise I'm fine with just doing the full freeze on the 15th.
15:49 < mdomsch> everyone's got the 29th free for an all-hands-on-deck, right?
15:50 < mmcgrath> mdomsch: I'll send a reminder out to everyone and about the change freeze.
15:50 < f13> mmcgrath: I think jsut full freeze on the 15th is fine, but I would like veto powers on any proposed changes leading up to then
15:50 < f13> well, I'd like releng to thave that, not me specifically
15:51 < mmcgrath> f13: hmm, I don't see any reason that wouldn't work. veto / revert powers. Plenty of changes go through without approval.
15:51 < f13> I think those two sentences contradict eachother
15:52 -!- giallu [n=giallu(a)81-174-46-50.dynamic.ngi.it] has joined #fedora-meeting
15:52 < mmcgrath> f13: they do, the fact is you can't have veto power on something that doesn't go through an approval process. But I'm happy to give it to you :)
15:52 < mmcgrath> f13: If you've got a problem with a change though bring it to the list, I'm sure we can all work something out.
15:52 < mdomsch> mmcgrath, one more - soc status on download1 routing?
15:53 < mmcgrath> mdomsch: you're guess is as good as mine on that one. They don't tell me anything / keep me up to date on anything.
15:53 < mdomsch> mgalgoci was going to stir up trouble on Monday morning
15:53 < mmcgrath> I can follow up in the ticket and let tom know.
15:53 < mdomsch> as it stands, we won't be able to feed our tier 1 mirrors, much less anyone else
15:53 < mmcgrath> mdomsch: do you know when we lost that link yet?
15:53 < mdomsch> weeks ago
15:53 < mmcgrath> I requested that information from the soc and heard nothing.
15:53 < mdomsch> I remember seeing a note at the time, thinking it would be fixed in a few days
15:54 < mmcgrath> ok. I'm going to bring it to my manager because A) its constant bs with the soc and I'm tired of it. and B) they never bothered to notify us.
15:54 < mdomsch> check mirror-list-d archives, they may have sent something there...
15:54 < mmcgrath> k, I'll double check there then.
15:54 < mdomsch> in any case, we've _got_ to get the bits to ibiblio, else we're scrod
15:54 < mmcgrath> That I'm sure we can do.
15:54 < mdomsch> (past pluperfect of screw)
15:54 < f13> mmcgrath: ok, so maybe the concept of a soft freeze is where you ask before making any changes, and at that point I'd have the ability to say no.
15:55 < f13> mmcgrath: at the hard freeze date, just no changes at all unless it's fixing things
15:55 < mmcgrath> f13: why not just stick with the system we put in place last release? It seemed to work just fine?
15:55 < f13> *shrug* sure, we could do that
15:55 * mmcgrath just doesn't want to complicate it unless there's a reason to. We've got a good team of people there.
15:56 < mmcgrath> Alrighty, anyone have anything else to discuss? If not we'll close in 30
15:56 < mmcgrath> 15
15:57 < f13> nod
15:57 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Meeting Closed!
16 years, 2 months
Change Freeze (IMPORTANT)
by Mike McGrath
Just a reminder to everyone that the change freeze will be taking place on
April 15th.
What does this mean? It means you don't make _ANY_ changes to our
infrastructure without running it by the list first with an explanation of
why it can't wait until after the freeze is over (April 30th)
These dates are dependent on if the release slips and such. Just like
last time this is a STRICTLY enforced rule.
If you've got any questions. Let us know.
-Mike
16 years, 2 months
service accounts for bz and FAS
by Matt Domsch
Per http://fedoraproject.org/wiki/MattDomsch/FTBFS, I created an
account in bugzilla, ftbfs(a)fp.o, which I will use for filing automated
bug reports. In aliases.template.erb I'm routing it's mail to
/dev/null. :-)
However, it dawns on me that someone _could_ come along and try to
create a username of "ftbfs" in FAS, which is undesirable.
Should I create a new account in FAS too, to "claim" it?
This account in bugzilla also needs to be in the bugzilla "editbugs"
group, in order to properly file bugs (you can't set bug_state,
blocked, or dependson fields via xmlrpc without being in editbugs).
I've asked dkl to grant editbugs to this account, but if I created the
account in FAS that would be another way to get that set. Except for
the whole "signing of the CLA" for what isn't a real person's account,
it's a "service" account. Which strikes me as odd.
Advice on the best way to handle this, and future similar account
needs?
Thanks,
Matt
--
Matt Domsch
Linux Technology Strategist, Dell Office of the CTO
linux.dell.com & www.dell.com/linux
16 years, 2 months
Looking for team reps
by Mike McGrath
I'm looking for team reps from Ambassadors and translators for the wiki
migration. Any volunteers?
-Mike
16 years, 2 months
xen7 networking issue?
by Toshio Kuratomi
Someone just reported that the wiki was down and I thought xen7 might
have crashed again. when I tried to ssh to it my connection timed out.
ping xen7 was fine. I serial consoled in and then immediately tried
to ssh to xen7 again. it worked. uptime showed that xen7 has been up
since this morning and the app servers are all running, etc. There are
no iscsi errors in /var/log/messages.
Is there a possibility that we're experiencing some sort of networking
issue with xen7? Maybe that issue is exacerbating the iscsi bug that
causes our xen hosts to crash?
-Toshio
16 years, 2 months
