#emc-devel | Logs for 2008-01-24

[00:56:55] <skunkworks633> skunkworks633 is now known as skunkworks
[04:14:49] <fenn> one problem with a graphical hal config tool is how to get things set up in the right order.. should this burden fall on the users?
[04:20:26] <jmkasunich> you mean the order of function calls in threads?
[04:22:41] <fenn> yes and also GUI stuff like scaragui (although that really should be done with NML imho)
[04:23:10] <fenn> <- distracted easily
[04:23:35] <jmkasunich> thats ok - I'm about to go to bed anyway
[04:24:03] <fenn> maybe it was the four bowls of cocoa crunchies
[04:24:38] <jmkasunich> gonna go to bed about 6 in the morning eh?
[04:24:38] <SWPadnos> heh
[04:25:08] <fenn> i gotta finish (start) this adapter thingie
[13:57:06] <skunkworks168> skunkworks168 is now known as skunkworks_
[15:38:01] <cradek> jepler: when do you expect to do 2.2.3?
[15:42:43] <jepler> cradek: I could do it this weekend if nobody's actively working on any bugs
[15:43:32] <jepler> I wish we could get a handle on that arc bug and fix it too
[15:46:50] <SWPadnos> hmmm
[15:47:02] <SWPadnos> #reqs%byteslast timefile
[15:47:03] <SWPadnos> 5581286.52%Jan/24/08 2:41 AM/iso/emc2.2.2-1-ubuntu6.06-desktop-i386.iso
[15:47:17] <SWPadnos> seems like a lot of requests for that iso, doesn't it?
[15:51:17] <jepler> how many full iso downloads does 86% represent
[15:51:50] <jepler> istm that some antisocial downloader programs work by starting many downloads at many different offsets, and even wget will resume an interrupted download
[15:51:53] <SWPadnos> well, the strange part is that there are more failed requests than there are requests
[15:52:24] <SWPadnos> #reqsfile
[15:52:26] <SWPadnos> 62803/iso/emc2.2.2-1-ubuntu6.06-desktop-i386.iso
[15:52:32] <SWPadnos> that's failed requests
[15:53:01] <SWPadnos> I noticed that the total bandwidth estimate is ~840GB this month, which seems a little high
[15:53:51] <jepler> I wonder what failure codes -- seems like that URL should never be 404'd, for instance
[15:54:46] <SWPadnos> http://linuxcnc.org/stats/
[15:54:51] <SWPadnos> use the emcboard login
[15:55:11] <SWPadnos> there's a separate failure code report, but it doesn't correlate to the individual files
[15:59:53] <jepler> it must be '503's
[16:00:38] <SWPadnos> oh good point - there are a lot of those
[16:01:04] <SWPadnos> and partial content probably
[16:01:25] <SWPadnos> I wonder ifcancelled downloads end up as partial content errors
[16:02:43] <jepler> 'partial content' isn't an error
[16:03:09] <jepler> the client can request byte ranges; the response to that request is '206 Partial Content' instead of '200 OK'
[16:03:57] <SWPadnos> ok
[16:06:11] <jepler> I wonder if there's some kind of rate limiting built into dreamhost
[16:06:48] <jepler> I wrote a program to repeatedly request that url 100 times .. it gives streaks of "200"s then streaks of "503"s
[16:06:53] <jepler> http://pastebin.ca/870969
[16:07:07] <SWPadnos> hmmm
[16:07:13] <SWPadnos> there shouldn't be
[16:07:31] <SWPadnos> the transfer limit on my account is now over 8TB/month
[16:07:59] <SWPadnos> there's no way that can be satisfied if they limit downloads (to one IP, on a cable / DSL connection anyway)
[16:08:11] <SWPadnos> there could be some DOS prevention thing though
[16:08:20] <jepler> this is issuing a lot of requests, not doing a lot of transfer
[16:09:06] <SWPadnos> right - that would be more DOS-ish - using up connections or whatever
[16:10:15] <SWPadnos> can you try that program again, but with a couple-second pause between requests? (or was there a pause before?)
[16:13:17] <alex_joni> 'lo guys
[16:13:40] <SWPadnos> hya
[16:13:42] <SWPadnos> +i
[16:15:49] <alex_joni> * alex_joni runs home
[16:15:52] <alex_joni> bbl
[16:16:02] <SWPadnos> see ya
[16:27:57] <jepler> SWPadnos: no pause, just the amount of time it took to get the response code back
[16:28:03] <SWPadnos> ok
[16:28:50] <jepler> SWPadnos: running now with a 2 second pause
[16:29:20] <SWPadnos> thanks - that was a curiously regular set of 200's then 503's
[16:30:48] <jepler> looks pretty similar with the 2-second pause
[16:30:56] <SWPadnos> ok, thanks
[16:31:30] <jepler> http://pastebin.ca/870988
[16:32:22] <jepler> Service Temporarily Unavailable
[16:32:23] <jepler> The server is temporarily unable to service your request due to maintenance downtime or capacity problems. Please try again later.
[16:32:37] <jepler> here's what I get if I load that same URL in my browser during one of the 503 clumps
[16:32:58] <SWPadnos> ok, interesting. I may ask them about that
[16:33:34] <jepler> I get the same if I load a different linuxcnc url
[18:24:52] <skunkworks_> what about this http://www.novell.com/products/realtime/eval.html ;)
[18:30:38] <alex_joni> 2.6.22 + preempt_RT
[18:42:52] <skunkworks_> Respond more rapidly to changing market conditions and business needs, and thus maximize your revenue opportunities
[18:43:00] <skunkworks_> uh huh
[18:57:47] <alex_joni> * alex_joni still waits for those revenue oportunities
[18:57:58] <SWPadnos> you didn't respond in real-time, sorry
[18:58:24] <SWPadnos> woohoo! time to eat
[18:59:01] <alex_joni> real-time eating?
[19:05:17] <skunkworks_> I think all eating is realtime.
[19:06:10] <alex_joni> how about digesting?
[19:06:15] <alex_joni> that's definately not realtime
[19:06:27] <skunkworks_> user space?
[19:06:43] <alex_joni> not even
[19:06:44] <skunkworks_> background application
[19:07:13] <alex_joni> digest &
[19:07:43] <alex_joni> nice digest &
[20:01:05] <cradek> jepler: thanks [again] for userland sim
[20:05:13] <jepler> cradek: welcome
[20:05:19] <jepler> how did it help you most recently?
[20:05:40] <cradek> well ... it didn't
[20:05:42] <cradek> but I still like it
[20:05:53] <cradek> the bug I fixed last didn't show up there for some reason
[20:06:34] <SWPadnos> heh
[20:06:52] <jepler> oh
[20:07:00] <jepler> you're still welcome