#emc-devel | Logs for 2008-07-01

[03:30:42] <seb_kuzminsky> oops
[13:19:50] <Roguish> good morning all. Little help here. please look at this: http://imagebin.org/21826
[13:21:05] <Roguish> Is this an error I'm creating or something more? if me, how? I get this using emc2-trunk on 3 different installations (3 unique computers with clean installs)
[13:22:12] <Roguish> oh, this is an 'AXIS error' 'unexpected realtime delay'
[13:22:36] <skunkworks_> few questions... Onboard video? what kind of hardware?
[13:23:03] <skunkworks_> could you run latency-test from terminal and see what numbers you get?
[13:23:35] <skunkworks_> also do a lspci -v and pastebin the results.
[13:23:55] <skunkworks_> also - what is your base period?
[13:24:50] <skunkworks_> This is also good reading http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?TroubleShooting#Unexpected_realtime_delay_check_dmesg_for_details
[13:29:26] <Roguish> good morning. yes, onboard vid. mb= asus p5gc-mx/133, 2g mem. http://www.asus.com/products.aspx?l1=3&l2=11&l3=498&model=1574&modelmenu=1
[13:31:13] <Roguish> lspci -> http://pastebin.com/m76adeb8
[13:32:41] <Roguish> here's a clean dmesg: http://pastebin.com/d3964b97d
[13:33:44] <Roguish> here's the full .ini file: http://pastebin.com/da8985f8
[13:35:05] <SWPadnos> it's got a 3GHz CPU?
[13:35:25] <SWPadnos> ah - 2.66
[13:36:53] <Roguish> here's a latency-test: http://pastebin.com/d5706cb8e
[13:37:17] <skunkworks_> You have to run it longer.
[13:37:29] <Roguish> how long?
[13:37:38] <SWPadnos> at least a couple of minutes
[13:37:50] <SWPadnos> and you should be doing "stressful" things while it runs
[13:38:00] <Roguish> can i pipe the output to a file?
[13:38:08] <SWPadnos> run glxgears, open OpenOffice, drag windows around ...
[13:38:16] <skunkworks_> you can run latency-test (it is a gui)
[13:38:26] <SWPadnos> don't worry about that - just copy/paste from the terminal once you stop the test
[13:38:46] <Roguish> running.
[13:39:03] <SWPadnos> (select an area then middle-click to paste, in case you don't know that trick :) )
[13:39:43] <skunkworks_> SWPadnos: tab complete working with directorys is awsome. It was a duh moment.
[13:39:50] <SWPadnos> heh
[13:40:14] <SWPadnos> tab completion has been in the shell for far longer than it's been in any program I know (with the possible exception of emacs or vi)
[13:40:49] <Roguish> i understand about 'stressing', but this happens just sitting there with AXIS running and nothing else going on. also, it is occuring on 3 seperate boxes.
[13:41:01] <SWPadnos> yep - two separate things
[13:41:23] <SWPadnos> it's bad enough to trigger the RT delay error, let's see how bad it is overall
[13:42:15] <Roguish> where is glxgears?
[13:42:23] <SWPadnos> terminal ;)
[13:42:33] <SWPadnos> hit ctrl-shift-t for a new terminal tab
[13:43:16] <Roguish> runnign glxgears full screen. openoffice, several firefox........
[13:44:03] <skunkworks_> what is your ovl max?
[13:44:09] <skunkworks_> at the moment
[13:44:26] <Roguish> at this time: 16519
[13:44:32] <skunkworks_> interesting
[13:44:43] <skunkworks_> let it keep running
[13:44:50] <Roguish> by the way, this is 8.04
[13:44:58] <skunkworks_> drag windows around
[13:45:34] <Roguish> stays fixed at 16519
[13:46:08] <Roguish> if emc2 and axis were running, the error would have come up long ago.
[13:46:23] <skunkworks_> you're not using any usb keychain drives are you?
[13:46:33] <Roguish> yes.
[13:46:43] <skunkworks_> do you have it plugged in now?
[13:46:47] <Roguish> yes
[13:46:48] <skunkworks_> oh
[13:47:04] <Roguish> remove it now.
[13:47:17] <skunkworks_> remove it and plug it back in for grins
[13:47:37] <Roguish> back in now.
[13:48:14] <skunkworks_> still 16519?
[13:48:23] <Roguish> up to 17241 a bit ago, then with thumb drive back in up to 22892
[13:48:40] <Roguish> all that other stuff is still running.
[13:49:07] <skunkworks_> not too bad.. What happens when you open the drive?
[13:50:41] <SWPadnos> you can probably stop the test now and pastebin the last few lines of output
[13:50:56] <Roguish> opened the file, and opened gedit with a file on the drive, still at 22892
[13:51:43] <Roguish> http://pastebin.com/d201635ee
[13:52:09] <SWPadnos> just for grins, please run the EMC latency test
[13:52:23] <SWPadnos> I think it may be in the CNC menu, or latency-test if not
[13:52:53] <Roguish> how?
[13:53:07] <skunkworks_> open terminal and type latency-test
[13:53:09] <SWPadnos> type latency-test in a terminal, or run the latency test from the menu if it's there
[13:53:23] <skunkworks_> I don't think it is a menu item yet.
[13:53:32] <Roguish> running
[13:53:34] <SWPadnos> ok - I thought it might be on the 8.04 CD
[13:53:58] <SWPadnos> now do the same things as before: glxgears, openoffice, window dragging/resizing ...
[13:55:22] <Roguish> ok, 3 glxgears running, another firefox, the drive is still in......
[13:56:06] <Roguish> ok, ?
[13:56:23] <SWPadnos> sounds good to me
[13:56:27] <skunkworks_> base period max latency?
[13:57:05] <SWPadnos> it looks like you have a lot of stuff running for your machine (scope, classicladder, some custom HAL files ...), I wonder if the servo thread is actually using to much time sometimes
[13:57:19] <Roguish> 'Base thread' Max Interval = 51479
[13:57:53] <Roguish> max jitter= 27444
[13:57:57] <Roguish> of base thread
[13:58:23] <jepler> you could pastebin the result of 'halcmd show thread' and 'halcmd show pin *tmax'
[13:58:40] <SWPadnos> yes - after you get the error
[13:58:52] <SWPadnos> of course, close the latency-test before running EMC
[13:59:55] <Roguish> ok, started emc
[14:00:33] <Roguish> now, just sitting with out running gcode or anything, the error will show in a minute or 2.
[14:00:40] <Roguish> go it.
[14:02:19] <jepler> op -- 'halcmd show param *tmax' (not pin)
[14:02:55] <SWPadnos> and halcmd show thread
[14:07:25] <Roguish> ok, i'm thick.. i run show thread > halinfo.txt where's the file get created?
[14:07:43] <Roguish> run from in HAL configuration in emc
[14:07:54] <SWPadnos> in the current directory
[14:08:59] <Roguish> can
[14:09:07] <Roguish> cannot find it.
[14:09:29] <SWPadnos> ls should show it to you
[14:09:50] <SWPadnos> wait - did you do that fom within halcmd?
[14:09:53] <SWPadnos> from
[14:10:43] <Roguish> i guess. from the command window in 'hal configuration' in emc.
[14:10:59] <SWPadnos> go to a terminal, and type `halcmd show param *tmax`
[14:11:12] <SWPadnos> then `halcmd show thread`
[14:12:03] <skunkworks_> do you need to do a . scripts\emc-environment first? (he is running trunk)
[14:12:14] <jepler> I doubt that you can do ">" shell redirections in the 'hal configuration' program
[14:12:19] <SWPadnos> oh, for trunk you probably do need to do that
[14:12:24] <SWPadnos> nope
[14:12:47] <SWPadnos> I'm not sure that commands work all the time either (I think there was a problem with that at some point)
[14:16:52] <Roguish> ok, just did the '. scripts/emc-environment', and now emc won't start:
[14:17:03] <SWPadnos> uh
[14:17:10] <SWPadnos> how do you normally start EMC?
[14:18:07] <Roguish> in term, with scripts/emc
[14:18:11] <Roguish> insmod: error inserting '/usr/realtime-2.6.24-16-rtai/modules/rtai_hal.ko': -1 File exists
[14:18:36] <SWPadnos> are you running a run-in-place EMC, or installed?
[14:19:09] <cradek> you can't run emc at the same time as the latency test
[14:19:25] <Roguish> run-in-place
[14:19:36] <Roguish> no lat tests running
[14:19:47] <SWPadnos> so you open a terminal, cd /path/to/emc, then scripts/emc ?
[14:20:02] <Roguish> yes
[14:20:18] <SWPadnos> and this is on a system that was installed from an EMC2 liveCD?
[14:21:21] <Roguish> i've done a half dozen or more install lately, i'm not sure it this was the live cd or using the install script after a new 8.04 install.
[14:21:28] <Roguish> probably the latter.
[14:22:02] <SWPadnos> ok, but there's an installed EMC that you can get to from the Applications menu though, right?
[14:22:09] <Roguish> yes
[14:22:21] <SWPadnos> hmmm. that should cause problems then :)
[14:22:39] <Roguish> ????????????????
[14:22:42] <SWPadnos> when you compiled this TRUNK, did you do the `sudo make setuid` step?
[14:22:49] <Roguish> yes
[14:22:54] <SWPadnos> ok
[14:23:22] <Roguish> i pretty much blindly follow what's in the wiki.
[14:23:26] <SWPadnos> a system that has an installed EMC has path and environment settings that point to the installed EMC
[14:23:50] <SWPadnos> when you run in place, you need to replace those with settings that point to the run-in-place location
[14:24:07] <SWPadnos> if you don't do that, then some things load from the compiled version, and others load from the installed version
[14:25:26] <skunkworks_> I have rip and installed here on my laptop and don't have issues running rip using the . scripts/emc....
[14:25:36] <SWPadnos> assuming that you have no latency test or EMC (or RTAI latency test) running, you should be able to open a terminal, type `. /path/to/emc/scripts/emc-environment`, and then run the R-I-P emc or halcmd by just typing emc or halcmd
[14:26:23] <SWPadnos> ecripts/emc may work, but when you go to another terminal and run halcmd, you end up with the installed one (I think)
[14:26:44] <SWPadnos> there are times when that doesn't work at all (when the SHM_ID changes, for example)
[14:28:08] <Roguish> just did that: http://pastebin.com/d21c5585e
[14:28:23] <jepler> insmod: error inserting '/usr/realtime-2.6.24-16-rtai/modules/rtai_hal.ko': -1 File exists
[14:28:47] <jepler> you have a left-over latency test or something running, or something didn't clean up after itself when it exited.
[14:29:02] <Roguish> all that testing goof something up?
[14:29:05] <Roguish> guess so.
[14:32:01] <Roguish> boot time?
[14:32:11] <Roguish> kill process?????
[14:32:12] <jepler> if you're not familiar with how to use lsmod/rmmod to clean things up, a reboot is the easiest way to fix that problem.
[14:32:12] <SWPadnos> shouldn't be
[14:32:24] <SWPadnos> will halrun -U do the trick?
[14:32:43] <jepler> SWPadnos: dunno, maybe
[14:33:33] <Roguish> did that one, don't think it worked.
[14:33:50] <SWPadnos> could you pastebin the output of lsmod?
[14:33:59] <SWPadnos> then reboot if it's the easiest thing to do :)
[14:35:14] <Roguish> http://pastebin.com/d69e27b04
[14:35:45] <SWPadnos> yep - rtai_hal is there
[14:36:02] <SWPadnos> if you haven't rebooted, try `sudo rmmod rtai_hal`
[14:37:04] <Roguish> in use.....
[14:37:15] <SWPadnos> oh duh - I should have seen that
[14:37:20] <Roguish> boot time?
[14:37:43] <SWPadnos> well, could you pastebin the output of `ps ax` ?
[14:38:46] <Roguish> http://pastebin.com/d6eda83b4
[14:39:00] <Roguish> gotta love that pastebin.
[14:39:43] <SWPadnos> ok - I've got to ru. I think rebooting is the easiest thing to do at the moment
[14:39:47] <SWPadnos> run, not ru
[14:40:13] <Roguish> booting now.
[14:48:41] <Roguish> back for more
[16:16:56] <Roguish> SWPadnos, cradek, jepler: thanks for the help. i'm off to Yosemite to hike up Half Dome. back by Friday. I would like to take this up again then, or possibly Saturday.
[16:17:09] <jepler> have fun
[16:33:49] <alex_joni> hi all
[16:36:12] <skunkworks_> Hi alex
[16:41:51] <alex_joni> * alex_joni finished work :D
[16:52:29] <skunkworks_> nice :)
[16:54:00] <alex_joni> yeah..
[17:02:46] <seb_kuzminsky> alex_joni: hi, i noticed your emc branch on launcpad
[17:02:56] <seb_kuzminsky> have you been playing with it? how's that working out?
[17:14:01] <alex_joni> seb_kuzminsky: if you have an launchpad account, I can add you
[17:14:13] <alex_joni> I poked around, but mostly to reserve it for "possible" future use
[17:19:32] <seb_kuzminsky> i've been using bzr for my hostmot2 development, and i just added my main hm2 dev branch to lp
[17:19:50] <seb_kuzminsky> https://code.launchpad.net/~seb-highlab/emc/hostmot2
[17:24:24] <alex_joni> gotta run for a bit, will be back later
[17:26:05] <seb_kuzminsky> ttyl
[18:51:46] <fenn_> fenn_ is now known as fenn
[19:20:11] <alex_joni> seb_kuzminsky: back
[19:44:11] <BigJohnT> hi alex
[19:45:51] <alex_joni> hi BigJohnT
[19:46:25] <BigJohnT> is it very involved to add a file to the docs?
[19:50:33] <cradek> what kind of file?
[19:50:57] <BigJohnT> lyx document file
[19:51:22] <alex_joni> nope
[19:51:31] <BigJohnT> for hardware examples
[19:51:31] <alex_joni> you simply create it on your local dir
[19:51:40] <BigJohnT> did that part :)
[19:51:41] <alex_joni> then use cvs add file/ cvs commit -m "message"
[19:51:57] <BigJohnT> ok easy enough
[19:51:58] <BigJohnT> thanks
[19:51:59] <alex_joni> but you need to be careful where you stick it
[19:52:08] <alex_joni> moving it around is a PITA with CVS
[19:52:36] <BigJohnT> I was thinking that the hardware examples needed a folder
[19:53:42] <cradek> is there some step to generate pdf/html etc from it?
[19:54:18] <cradek> well no worry, someone can help with that
[19:54:27] <cradek> I see you have to do something in docs/src/Submakefile
[19:54:53] <BigJohnT> ok
[19:55:08] <BigJohnT> I'm not at that computer at the moment so I can't look
[19:55:34] <BigJohnT> if you add a part to the integrator manual will it not create the pdf/html?
[19:56:16] <cradek> I'm not sure - it looks a bit intricate
[19:56:35] <cradek> ask for help if it doesn't do what you want :-)
[19:56:51] <alex_joni> I think it does
[19:57:08] <alex_joni> there's a python foo jepler wrote which checks dependencies from lyx included files
[19:57:17] <BigJohnT> ok, it will be later in the week I think
[19:57:23] <BigJohnT> before it is ready
[19:57:25] <jepler> the listing in the makefile is for the html, I think
[19:57:30] <alex_joni> for html you have to add some foo in Submakefile, but someone else can do it too
[19:57:33] <jepler> the html doesn't look at the "container" lyx files
[19:57:40] <alex_joni> jepler: indeed it does :)
[19:57:45] <alex_joni> s/does/is/
[19:57:58] <alex_joni> must be true if we both remember it the same way :D
[19:59:10] <seb_kuzminsky> alex_joni: hi!
[20:00:48] <alex_joni> hi
[20:05:05] <jepler> I wonder if 2.3 should stop pretending support for pre-2.6 kernels; I don't believe we regularly test on any or get any user reports about old kernels either.
[20:06:22] <seb_kuzminsky> hostmot2 wont work on pre-2.6 kernels as is
[20:06:31] <seb_kuzminsky> i'm for dropping 2.4
[20:10:30] <alex_joni> hmm.. haven't tested 2.4 in a while either
[20:11:23] <seb_kuzminsky> anything not tested by the build farm should be considered "unmaintained, unsupported, suspected broken"
[20:12:48] <jepler> we only know whether stuff on the build farm builds, not whether it runs :-P
[20:13:27] <cradek> I tested a 2.4 build not long ago - it worked.
[20:13:55] <cradek> but tcl, python, etc., were all too old. pretty much just the kernel stuff worked. I only got the basic emc/xemc to run.
[20:16:22] <seb_kuzminsky> jepler: i'd like to work on unit testing and integration testing of emc2 one day... and roll it into the build system
[20:20:54] <jepler> seb_kuzminsky: cool
[20:21:14] <jepler> cradek: great, so you're the volunteer to maintain 2.4 support now
[20:21:34] <seb_kuzminsky> d'oh cradek!
[20:24:04] <cradek> haha
[20:25:07] <jepler> cradek: have any kernel 2.2 systems in the closet?
[20:25:14] <cradek> I can take or leave 2.4. If it's working we shouldn't tear it out, but as soon as it's a pain (and nobody steps up to maintain it), whatever.
[20:25:39] <cradek> jepler: no, the first machine I ran emc on was 2.4.
[20:37:04] <jepler> linux 2.6.0 was released on 2003-12-18. fedora core 2 (2004) and
[20:37:06] <jepler> d
[20:37:18] <jepler> newer, and every ubuntu version ever, use it
[20:37:33] <seb_kuzminsky> btw, where is the buildfarm webpages?
[20:37:44] <jepler> debian wated until 2005 for it
[20:38:02] <jepler> seb_kuzminsky: linuxcnc.org/c_f or /compile_farm if you prefer the joomla decorations on the page
[20:38:14] <seb_kuzminsky> thx
[20:47:15] <alex_joni> * alex_joni heads to bed
[20:47:19] <alex_joni> good night all
[20:47:26] <seb_kuzminsky> good night alex
[20:51:29] <seb_kuzminsky> the dapper buildslave hasnt run since april