#emc-devel | Logs for 2007-12-05

[00:51:23] <CIA-19> EMC: 03cradek 07TRUNK * 10emc2/src/emc/rs274ngc/ (8 files): G73 drill cycle with chip breaking
[13:31:15] <skunkworks_> Thanks :)
[13:31:19] <alex_joni> skunkworks_: np
[13:31:28] <alex_joni> let me know how it looks after you install
[13:31:41] <skunkworks_> 36%
[13:32:02] <alex_joni> of the install?
[13:32:04] <skunkworks_> yes
[13:32:15] <alex_joni> you need a 16+ core machine
[13:32:32] <skunkworks_> heh - this is a III 550 I think
[13:32:46] <SWPadnos> http://server1.streamsend.com/streamsend/clicktracker.php?cd=3854&ld=8&md=34&ud=06b6b0fd2a02ec78ccc7744e748f0c48&url=http://www.aberdeeninc.com/abcatg/Stonehaven-A135.htm
[13:32:54] <skunkworks_> seems to run emc quite nice.
[13:33:03] <SWPadnos> or just http://www.aberdeeninc.com/abcatg/Stonehaven-A135.htm
[13:33:37] <skunkworks_> my coworker (around 22 years old) has a bunch of amd stock. He has lost a ton of money so far ;)(
[13:34:33] <skunkworks_> He had decided to just let it ride and hope for the best.
[13:36:12] <SWPadnos> yay - only 550 more emails to wade through
[13:36:31] <skunkworks_> yikes. 1% good?
[13:36:32] <alex_joni> * alex_joni puts http://www.aberdeeninc.com/abcatg/Stonehaven-X840.htm on the christmas wishlist
[13:37:07] <SWPadnos> no, this is after I've culled most of the spam
[13:37:19] <skunkworks_> wow - your a popular guy.
[13:37:29] <SWPadnos> heh - sure, that's it
[13:37:43] <SWPadnos> I'm on several mailing lists - that's the other possible explanation ;)
[13:40:17] <jepler> wow only $8313 for quad quad 2.0GHz, 16GB? Send me one too!
[13:40:33] <jepler> http://xkcd.com/353/
[13:42:58] <alex_joni> jepler: lol
[13:44:34] <jepler> cradek: you'll have to help me document g73
[13:51:31] <CIA-19> EMC: 03tissf 07TRUNK * 10emc2/docs/src/hal/halshow_fr.lyx: French translation typo!
[13:52:05] <skunkworks_> alex_joni: it uses your desktop piture.
[13:52:14] <skunkworks_> (penguin and 2.2)
[13:53:23] <CIA-19> EMC: 03tissf 07v2_2_branch * 10emc2/docs/src/ (Master_HAL_fr.lyx Submakefile index.tmpl index_fr.tmpl): French translation
[13:53:23] <CIA-19> EMC: 03tissf 07v2_2_branch * 10emc2/docs/src/gui/axis_fr.lyx: French translation
[13:53:24] <CIA-19> EMC: 03tissf 07v2_2_branch * 10emc2/docs/src/config/ (copy_and_run_fr.lyx ini_homing_fr.lyx): French translation
[13:53:24] <CIA-19> EMC: 03tissf 07v2_2_branch * 10emc2/docs/src/hal/halshow_fr.lyx: French translation
[13:54:27] <alex_joni> skunkworks_: it's too smart :)
[13:54:36] <skunkworks_> heh :)
[14:00:03] <alex_joni> skunkworks_: I had help though: http://icanhascheezburger.files.wordpress.com/2007/02/leveledup.jpg
[14:02:05] <alex_joni> http://icanhascheezburger.files.wordpress.com/2007/02/nightout.jpg (seems familiar)
[14:06:32] <jepler> somebody needs to make an emc control panel with these buttons: http://www.sparkfun.com/commerce/product_info.php?products_id=8493
[14:10:47] <SWPadnos> wow - I'm surprised (a little) that those are still $50 each
[14:11:56] <jepler> yeah that's one thing that is stopping me from doing it myself
[14:17:00] <Guest404> Guest404 is now known as skunkworks__
[14:29:39] <alex_joni> bbl
[14:48:24] <cradek> I wonder if anyone is using the screw mapping
[14:48:50] <jepler> I thought it was contributed, in which case somebody must be
[14:49:02] <jepler> or am I thinking of the velocity-limiting for backlash?
[14:49:26] <skunkworks__> I thought jmkasunich had gotten it working the last time. (about all I remember)
[14:49:28] <cradek> I think jmk wrote screw comp
[14:50:31] <jepler> cradek: oh by the way -- for new error messages in the interpreter, I believe it's better not to define new error numbers but to use the newer error macros that let you specify the string in-line
[14:51:04] <cradek> that's good, because I'm sick of doing them that way
[14:51:17] <cradek> but, is it harder for i18n?
[14:51:25] <jepler> no, it should work about the same for i18n
[14:52:06] <cradek> ok. when I merge g17.1 etc, there will be a lot of new ones. They already will clash, so I'll fix them that way
[15:29:27] <CIA-19> EMC: 03tissf 07v2_2_branch * 10emc2/docs/src/hal/ (general_ref_fr.lyx halshow_fr.lyx tools_fr.lyx): French translation update
[15:32:17] <CIA-19> EMC: 03tissf 07TRUNK * 10emc2/docs/src/hal/ (general_ref_fr.lyx halshow_fr.lyx tools_fr.lyx): French translation update
[16:06:16] <skunkworks__> alex_joni: banging on the installed emc 2.2 - no issues so far.
[16:13:58] <cradek> jepler: why don't we change all the errors to the new format?
[16:14:28] <jepler> cradek: I don't see a problem with doing that in 2.3.
[16:16:09] <jepler> cradek: what is the name of the G73 canned cycle?
[16:17:05] <cradek> how about 'drill cycle with chip breaking'
[16:18:24] <jepler> OK that's what I had come up with too
[16:19:34] <jepler> http://pastebin.com/m646ae0ed
[16:20:10] <jepler> er, must be "steps 2 and 3"
[16:20:55] <cradek> I don't like the use of "down" and "out" as opposites
[16:21:00] <cradek> I recommend using "up"
[16:21:32] <jepler> I came up with that because G83 said "backed off a bit"
[16:21:37] <jepler> -> "back off"
[16:21:51] <jepler> 3. Rapid up a bit
[16:22:09] <cradek> ... to break the chip
[16:22:10] <cradek> ?
[16:22:30] <cradek> maybe not, it's fine without that
[16:22:44] <jepler> it is not typical for the purpose of each step to be explained
[16:22:49] <cradek> ok
[16:23:05] <cradek> looks good then, thank you
[16:23:55] <CIA-19> EMC: 03jepler 07TRUNK * 10emc2/docs/src/gcode/main.lyx: document new canned cycle G73
[16:24:34] <CIA-19> EMC: 03jepler 07TRUNK * 10emc2/docs/html/gcode.html: mention new G76 canned cycle
[16:25:47] <jepler> er, that commit message is wrong but I think I got it right in the document
[16:26:19] <jepler> cradek: what link from the gcode quick ref did you say did not work? Were you referring to 2.2 or to TRUNK?
[16:26:29] <cradek> it was 2.2
[16:26:30] <cradek> I think G18
[16:26:56] <jepler> OK thanks
[16:28:11] <CIA-19> EMC: 03jepler 07TRUNK * 10emc2/docs/html/gcode.html: fix G18, G19 links to full docs
[16:30:36] <CIA-19> EMC: 03jepler 07v2_2_branch * 10emc2/debian/changelog: gcode quick reference had wrong links to G18, G19 documentation
[16:30:36] <CIA-19> EMC: 03jepler 07v2_2_branch * 10emc2/docs/html/gcode.html: gcode quick reference had wrong links to G18, G19 documentation
[16:32:20] <cradek> cool
[16:47:51] <CIA-19> EMC: 03jepler 07TRUNK * 10emc2/docs/src/ (checklinks.py Submakefile): automated check that links in quickref work
[16:50:11] <jepler> .. because manually clicking each link is STUPID
[16:50:28] <SWPadnos> heh
[16:51:01] <jepler> (though I suppose the program can only verify that the anchors exist, not that they are the right ones)
[17:02:40] <jepler> argh why am I laptop-shopping?
[17:06:44] <cradek> in spring I always start looking at cars. maybe you have a seasonal problem like that
[17:08:17] <jepler> I seem to computer shop during all seasons
[17:08:38] <SWPadnos> its a pan-seasonal problem
[17:09:54] <SWPadnos> I just bought another laptop
[17:10:02] <SWPadnos> http://www.laptopgiving.org/en/index.php
[17:10:39] <jepler> that's something I'm trying carefully not to be interested in
[17:10:44] <SWPadnos> heh
[17:11:49] <jepler> also the keyboard looks just terrible to use
[17:11:54] <SWPadnos> true
[17:12:21] <SWPadnos> I figured it would be nice for some kid soemwhere, plus maybe another kid (or older person) here, if I give away the one they send me
[17:13:45] <jepler> the placement of the headphone jack is very interesting
[17:13:59] <jepler> so much for trying carefully not to be interested..
[17:14:08] <SWPadnos> oops
[21:21:53] <jepler> (I think the latest versions of gcam require a gtk+ version not available on dapper?
[21:21:57] <jepler> )
[21:22:04] <jepler> (which really makes it pretty useless to emc users)
[21:22:20] <cradek> hmm.
[21:23:16] <jepler> GCAM requires the following dependencies in order for the "configure" script to succeed.
[21:23:19] <jepler> * GTK+ 2.10.0 or higher
[21:23:25] <jepler> $ dpkg -l libgtk2.0-dev
[21:23:27] <jepler> ii libgtk2.0-dev 2.8.20-0ubuntu Development files for the GTK+ library
[21:23:50] <cradek> inexorable progress and all that
[21:24:22] <jepler> I have mentioned this to him before and actually submitted a patch at the time, but apparently he's not that interested in compatability with the systems where emc runs
[21:24:34] <cradek> I thought he fixed it.
[21:24:37] <cradek> he must have changed his mind back.
[21:29:07] <jepler> probably ran into some other feature in 2.10 that was desirable. I wish it was on dapper, because then I would have used GtkAssistant as the 'wizard' widget in stepconf; as it was, I had to require Gnome for GnomeDruid instead.
[21:29:59] <lerneaen_hydra> jepler, it require gnome-desktop?
[21:31:03] <jepler> lerneaen_hydra: specifically, it requires python2.4-gnome2
[21:32:22] <lerneaen_hydra> oh, is it gnome specific or relatively common? (just curious)
[21:34:22] <jepler> if the packager took care when splitting up gnome into its constitutent packages, it should not be necessary to pull in any gnome desktop applications, but just a few core libraries. However, I don't know whether the packaging was done to enable this or not
[21:34:49] <lerneaen_hydra> jepler, ah, I see
[21:34:58] <jepler> it won't matter to the typical emc user, since they get a full gnome desktop when installing from the emc live cd anyway
[21:35:59] <lerneaen_hydra> most installations have the gnome desktop?
[21:36:02] <jepler> oh, and classicladder pulls in libgnomeprintui2.2 to enable printing of ladders
[21:36:14] <jepler> if you install from the emc2 live cd or from the standard ubuntu cd then you get gnome
[21:36:17] <lerneaen_hydra> oh ok
[21:36:24] <jepler> if you install xubuntu or kubuntu then you don't have it by default
[21:36:28] <lerneaen_hydra> yeah but do most people use that installation method?
[21:36:47] <lerneaen_hydra> instead of using some other distro or something else a bit more exotic
[21:37:04] <jepler> no most people show up and say "I never used linux before will you compile a real time kernel on my gentoo slackware suse"
[21:37:10] <jepler> but they never actually get emc to run
[21:37:22] <lerneaen_hydra> say a company that boxes an emc2 installation for example (have you heard of companies using emc?)
[21:37:34] <lerneaen_hydra> ugh, that's not what I was thinking of ;)
[21:40:15] <jepler> I don't know the answer to your question.
[21:41:20] <jepler> since it is impossible to buy a new hard disk so small that it can't hold the regular ubuntu desktop -- and getting harder every month to buy a piece of flash memory without that much space -- I won't lose much sleep over the possibility that this package requirement is a hardship for anyone
[21:41:41] <lerneaen_hydra> oh I wasn't implying that it was a problem
[21:42:22] <lerneaen_hydra> just a bit, well, ugly ;) if you know what I mean
[21:44:47] <fenn> it goes against my religion
[21:44:58] <fenn> i've sworn off gnome packages
[21:45:16] <jepler> as they say, if your eye offends you, pluck it out
[21:48:07] <jepler> bbl
[21:48:20] <alex_joni> fenn: we can even help with that
[21:49:25] <lerneaen_hydra> * lerneaen_hydra hands alex_joni a scalpel
[21:52:26] <alex_joni> nah.. to clean
[21:52:30] <alex_joni> too clean
[22:00:26] <SWPadnos> a shop-vac with the itty-bitty end would do it
[22:01:10] <skunkworks__> finally got this stupid raid 5 array to rebuild. Only been a week or so.
[22:01:20] <SWPadnos> you need faster disks
[22:01:27] <skunkworks__> 3%
[22:02:06] <skunkworks__> no - it was a bunch of issues.. The offical seagate replacements for the maxtor drives that where originally in the server where just a tad smaller.
[22:02:16] <SWPadnos> oops
[22:02:18] <skunkworks__> yah
[22:02:22] <SWPadnos> a tad larger is the way to go
[22:02:48] <skunkworks__> so they sent me some tad bigger fugitus drives - which also ended up being a tad smaller.
[22:04:26] <skunkworks__> so I ended up buying some seagate 140gb drive. (original was 73gb)
[22:06:14] <skunkworks__> Then I could set it as a hot spare - but it would not rebuild becuase there was a fantom drive in the array for some reason. Finally gave up (these are not hot swapable) and pulled the power out of the new drive - changed it to the id of the fantom drive. plugged it back in and boom - started rebuilding. pain
[22:07:33] <SWPadnos> and 12 minutes later, it's done! :)
[22:07:54] <skunkworks__> heh - it is going to be a bit. 5% so far.
[22:07:58] <skunkworks__> 7 drives
[22:08:11] <SWPadnos> ok, 7@73GB is a fair amount to rebuild
[22:08:15] <skunkworks__> yes
[22:08:25] <skunkworks__> almost a 1/2 TB
[22:08:38] <SWPadnos> slightly more, actually
[22:08:44] <SWPadnos> 490+21GB
[22:08:48] <SWPadnos> =511
[22:08:55] <SWPadnos> including parity though
[22:09:09] <skunkworks__> no - raid 5 6*73=438
[22:09:25] <SWPadnos> oh, is #7 a hot spare?
[22:09:49] <skunkworks__> no - but you loose the capacity of 1 of the drives in a raid 5 setup
[22:09:54] <alex_joni> #7 is parity
[22:10:04] <alex_joni> actually distributed parity.. but as a concept
[22:10:06] <skunkworks__> well - spread over all drives
[22:10:07] <skunkworks__> right
[22:10:10] <SWPadnos> sure, that's why I said "including parity though"
[22:10:39] <skunkworks__> Riiight :) usable space though is less than .5TB
[22:10:42] <SWPadnos> since the parity info also has to be regenerated/restored, the RAID needs to write out the raw capacity of all drives, not the usable space
[22:10:46] <SWPadnos> agreed
[22:11:03] <SWPadnos> is this SCSI or a bunch of raptors?
[22:11:12] <skunkworks__> scsi
[22:11:32] <skunkworks__> 320uw
[22:11:33] <SWPadnos> ok - those must be somewhat expensive replacement drives then :)
[22:11:37] <skunkworks__> yes
[22:11:49] <skunkworks__> that is why I wanted to stick with 73gb drives
[22:11:51] <SWPadnos> heh
[22:11:55] <skunkworks__> :)
[22:12:22] <SWPadnos> I finally broke down and bought a SATA raid for my server. I'd prefer SCSI, but for the workload, the price wasn't justifiable
[22:13:03] <alex_joni> got a link?
[22:13:24] <SWPadnos> err - not really. it's a 3Ware card and cage from NewEgg
[22:13:52] <SWPadnos> 4x300GB drives = 600GB usable + parity (RAID 5) + hot spare
[22:14:17] <skunkworks__> I just bought one of these - and 6 750GB sata drives.. Let you know how it works out.
[22:14:19] <skunkworks__> http://www.meritline.com/5-bay-esata-raid-hard-drive-storage-array.html
[22:14:20] <SWPadnos> or maybe they're 320GB, I don't remember
[22:14:39] <SWPadnos> sounds like you may have one too many drives ;)
[22:14:53] <skunkworks__> I know now to buy one extra at the time :)
[22:14:59] <SWPadnos> heh
[22:15:10] <skunkworks__> they where cheap - they all came to $800
[22:15:27] <SWPadnos> that's relatively amazing, isn't it?
[22:15:32] <skunkworks__> and seagate.
[22:15:42] <skunkworks__> it was a deal from I think buy.com
[22:15:59] <skunkworks__> could not pass it up
[22:16:16] <SWPadnos> $800 including that RAID cage?
[22:16:43] <skunkworks__> nop
[22:16:45] <skunkworks__> no
[22:16:51] <skunkworks__> just the drives
[22:16:59] <SWPadnos> ah, ok. still a good deal
[22:17:01] <skunkworks__> raid cage was 500 shipped I think
[22:17:17] <skunkworks__> drives wise - yes - cage - I don't really know
[22:17:22] <SWPadnos> $800 is what the computer club paid for a 10M external disk when I was in high school
[22:17:26] <skunkworks__> I wanted esata for the server.
[22:17:30] <skunkworks__> heeh
[22:17:48] <SWPadnos> it could have been $895 though, I don't remember
[22:17:49] <skunkworks__> yeh - i remember being excited about $1 a megabyte.
[22:17:55] <SWPadnos> yeah
[22:18:20] <SWPadnos> a friend and I pooled our money to buy a hard drive wehn they first hit 1G - ir was about $500
[22:18:25] <SWPadnos> it
[22:18:37] <SWPadnos> same deal for a 16M SIMM ($495)
[22:25:42] <skunkworks__> 6 $149.99 $899.94 free shipping
[22:26:00] <skunkworks__> looks like it has expired
[22:26:02] <skunkworks__> http://www.geeks.com:80/details.asp?invtid=ST3750640AS-R&cm_mmc=geekmail-_-daily_html-_-20nov07_750SEA-_-750SEAmain
[22:26:22] <skunkworks__> ok - 900 dollars ;)
[22:26:26] <SWPadnos> no bigie - that's only $10 less than NewEgg is now for a WD drive
[22:26:29] <SWPadnos> biggie
[22:27:03] <skunkworks__> I really have had bad luck with WD drives though.. If a drive is going to go bad - it seems to always be a WD
[22:27:26] <SWPadnos> I think I know at least one person who has had that problem with every drive manufacturer
[22:27:33] <skunkworks__> I bet.
[22:27:43] <SWPadnos> except Fujitsu (that I can recall)
[22:28:00] <SWPadnos> but I don't know many people who use Fujitsus, so that's not a good data point
[22:28:34] <skunkworks__> heh - that is what is in our novell sever..
[22:28:43] <skunkworks__> 4 30somethingGB drives
[22:29:49] <skunkworks__> they have all been replaced - they where originally IBM drives made by fugitu. I have replaced every one with the same fugitsu drives now. BUt this server has got to be atleast 8 years old
[22:31:13] <alex_joni> there are special datacenter editions from all HDD manufacturers
[22:31:18] <alex_joni> tuned to work 24/7
[22:31:50] <alex_joni> I think it's ES series from WD
[22:32:33] <skunkworks__> all the WD i have had gone bad have been in cheap workstations.
[22:33:35] <skunkworks__> alex_joni: the live cd looks good.. No issues so far
[22:34:46] <skunkworks__> although I have not actaully run any hardware with it.
[22:34:49] <alex_joni> skunkworks__: good.. I changed it on the webpages
[22:35:03] <skunkworks__> sweet.
[22:35:05] <alex_joni> skunkworks__: that test is not for the Live CD.. but for emc2.2.2
[22:35:15] <alex_joni> which afaik works a treat so far ;)
[22:49:43] <alex_joni> SWPadnos: how does this look? http://www.supermicro.com/products/chassis/2U/823/SC823S-R500LP.cfm
[22:50:43] <skunkworks__> * skunkworks__ love supermicro equipent..
[22:50:50] <skunkworks__> yikes
[22:51:03] <skunkworks__> * skunkworks__ loves supermicro equipment
[22:51:17] <alex_joni> that can be interpreted :D
[22:51:30] <alex_joni> is 2U enough for regular PCI cards?
[22:53:10] <alex_joni> or should I go with an 3U or even 4U ?
[22:54:21] <skunkworks__> I have not bought any rack mount servers..
[22:54:49] <alex_joni> next you'll be telling me you run windows on them too
[22:55:16] <skunkworks__> lets see - windows 2000, novell 4.11, and redhat.
[22:55:34] <skunkworks__> I think that sums it up. - actually we have a server coming in running 2003 now too.
[22:55:37] <alex_joni> :)
[22:55:52] <skunkworks__> hence why I am still at work
[22:56:09] <alex_joni> are SAS and SATA using the same hotplug bays?
[22:56:58] <skunkworks__> doesn't look like it..
[22:57:15] <skunkworks__> wait - hold on
[22:57:41] <skunkworks__> well Ill be - looks like it does.
[22:58:14] <skunkworks__> I have a sas server (not running at the moment) and the back of the drives look identical to the sata drives.
[22:58:17] <skunkworks__> I did not know that.
[22:58:39] <alex_joni> The SAS connector is a universal interconnection that is form-factor compatible with SATA. It allows SAS or SATA drives to plug directly into a SAS environment for mission-critical applications with high-availability and high-performance requirements or lower-cost-per-gigabyte applications such as near-box storage.
[22:59:34] <alex_joni> seems you can even run SAS disks on SATA controllers
[22:59:39] <skunkworks__> scary
[22:59:44] <alex_joni> http://www.digit-life.com/articles2/sas-and-sata/
[22:59:48] <skunkworks__> that just doesn't seem right ;)
[23:00:39] <alex_joni> sure does to me
[23:00:55] <alex_joni> it means cheaper bays, backplanes, cables, etc
[23:02:26] <SWPadnos> alex_joni, that looks like a nice server chassis, though I have no experience with it
[23:06:08] <alex_joni> SWPadnos: but generally.. is 2U a big enough size?
[23:06:23] <alex_joni> the ones I have now are all 4U (don't quite recall how high 2U is)
[23:07:23] <SWPadnos> 2u is ~3.5"
[23:12:25] <alex_joni> good night all
[23:14:07] <SWPadnos> night Alex