IRC log for #uphpu on 20090318

00:23.10*** join/#uphpu mastergunns1 (n=mastergu@c-71-199-7-162.hsd1.ut.comcast.net)
01:15.53*** part/#uphpu mastergunns1 (n=mastergu@c-71-199-7-162.hsd1.ut.comcast.net)
02:46.47*** join/#uphpu undertakingyou (n=will@undertakingyou.dsl.xmission.com)
02:46.47*** mode/#UPHPU [+v undertakingyou] by ChanServ
03:03.34*** join/#uphpu beandog (n=sdibb@gentoo/developer/beandog)
03:03.34*** mode/#UPHPU [+v beandog] by ChanServ
03:05.22beandogfun with sql?
03:08.39DexterTheDragonyeah ya know, like fun with electricity
03:10.48beandogfun with the dentist
03:11.15DexterTheDragonyeah things of that nature
03:17.14*** join/#uphpu mgeary (n=mgeary@69.169.142.45.provo.static.broadweave.net)
03:17.18*** mode/#UPHPU [+v mgeary] by ChanServ
03:27.50beandogI think Im gonna go look for DVDs
03:49.15*** join/#uphpu carmony (n=justinca@c-67-164-249-224.hsd1.ut.comcast.net)
03:49.15*** mode/#UPHPU [+v carmony] by ChanServ
04:20.01*** join/#uphpu mrkrow (n=chad@c-76-27-25-249.hsd1.ut.comcast.net)
04:20.01*** mode/#UPHPU [+v mrkrow] by ChanServ
06:31.24*** part/#uphpu mgeary (n=mgeary@69.169.142.45.provo.static.broadweave.net)
10:22.44*** join/#uphpu itimber (n=itimber@c-71-195-203-159.hsd1.ut.comcast.net)
10:22.44*** mode/#UPHPU [+v itimber] by ChanServ
13:43.20*** join/#uphpu Utah_Dave (n=David@72-166-51-162.dia.static.qwest.net)
13:43.20*** mode/#UPHPU [+o Utah_Dave] by ChanServ
14:18.57*** join/#uphpu utahcon (n=utahcon@bromine.sosstaffing.com)
14:39.16DexterTheDragonlivin the dream
14:49.18macnewboldthat's awesome
14:49.32macnewboldwhat's with all the human wimps who only send their ashes into space _after_ they're dead
14:50.52utahconSheep + LEDs = AWESEOME! http://i.gizmodo.com/5172963/sheep+powered-led-display-lights-up-welsh-hillside
14:57.43*** join/#uphpu carmony (n=justinca@c-67-164-249-224.hsd1.ut.comcast.net)
14:57.43*** mode/#UPHPU [+v carmony] by ChanServ
14:58.46*** join/#uphpu fungus (n=olsonl@bromine.sosstaffing.com)
14:58.46*** mode/#UPHPU [+o fungus] by ChanServ
14:59.34*** join/#uphpu SunSparc (n=sparky@173.8.91.177)
14:59.34*** mode/#UPHPU [+v SunSparc] by ChanServ
15:07.57*** join/#uphpu _psychic_ (n=_psychic@173.8.91.177)
15:07.57*** mode/#UPHPU [+v _psychic_] by ChanServ
15:17.01*** join/#uphpu mgeary (n=mgeary@166.70.44.49)
15:17.01*** mode/#UPHPU [+v mgeary] by ChanServ
15:19.02mgeary~yow
15:35.59*** join/#uphpu thebigdog (n=thebigdo@207.135.132.13)
15:35.59*** mode/#UPHPU [+o thebigdog] by ChanServ
15:40.46*** join/#uphpu mrkrow (n=chad@library.xwi.xmission.com)
15:40.46*** mode/#UPHPU [+v mrkrow] by ChanServ
15:40.51mindjujuso fungus, when you installed ubuntu on that sunfire, did you customize the CD at all?
15:41.02fungusNope
15:41.14mindjujuman, i'm so close on this, i know it!
15:41.14fungusJust standard install through serial console.
15:41.26mindjujuwait, what about the graphics though?
15:41.37fungusgraphics?  lol
15:41.41fungusnope
15:41.51mindjujuyeah, i don't want the graphics
15:42.04mindjujubut it seems to get stuck in this one part, like it's waiting for some kind of input
15:42.22mindjujui read somewhere that when boot comes up, i need to type linux console="ttyS0,9600"
15:42.30mindjujubut when i do that, it says i'm wrong
15:42.41mindjujubut if i hit 1, then it'll try to load off the CD
15:42.41fungusYeah, that's for x86 machines
15:42.55mindjujui've found out it seems to be reading off of silo.conf
15:45.21fungushrm, too long ago to remember anything like that.
15:46.01mindjujui've left it at the spot where it seems to be waiting for input, for about 30 min with no change
15:46.16mindjujuthe last line is something about Journaling or something, i should have it shortly
15:46.43mindjujuk, so i need to find what the right command for sparc is for linux console=ttyS0,9600
15:47.35fungusI don't think you need it at all.
15:47.50mindjujuthis is the last lines that appear
15:47.51mindjuju<PROTECTED>
15:47.52mindjujufsck.ext3: Unable to resolve 'UUID=b7a6d81d-7558-413b-bd85-4a65757bdedf'
15:47.52mindjuju/dev/sda1: recovering journal
15:47.59macnewboldnice
15:48.01mindjujuit'll stay here for more than 30 minutes
15:53.19utahconmissing a drive?
15:53.52*** join/#uphpu andrey (n=andrey@elphelut.fttp.xmission.com)
15:53.59mindjujuis that what that means?
15:55.39Dezertolok.. database guys.. what's the official on using "auto incrementing ids" I was told not to use them
15:55.57Dezertolby my database class teacher guy..
15:56.03mindjujui love 'em
15:56.12mindjujuwhat database though?
15:56.24Dezertolmysql
15:56.27fungusWhat were the reasons not to use them?  
15:56.41Dezertolaparently.. it's "bad design"
15:56.55Dezertolsomthing do with merging records
15:57.07Dezertoland restoring backups and stuff
15:57.26DezertolI've been using uniqid's
15:57.32Dezertolin both perl and php for my stuff
15:57.41Dezertolbut.. my lattest project..
15:57.52Dezertoli'm getting duplicat key violations
15:57.53mindjujui like auto increment because its much smaller than uuids
15:58.16mindjujuwait wait, you can't get duplicate keys with a unique id
15:58.27Dezertolyou spin them fast enough you can
15:58.29mindjujui think the world is suppose to explode if you do
15:58.36_carmonyif you have to merge records, than GUIDS typically are better than auto ids.
15:58.49DezertolI've got 3 scripts running at the same time importing over 1 mil each
15:58.54*** join/#uphpu josephscott (n=josephsc@75-169-255-227.slkc.qwest.net)
15:58.54*** mode/#UPHPU [+v josephscott] by ChanServ
15:59.05Dezertolthe 3rd script died due to dup key's
15:59.27_carmonywell, that sounds like a unique situation where GUIDs should be used
15:59.28mgearyi like auto increments, but i know that many dbs don't offer them
15:59.59Dezertolthe guy that taught my db class was an oracle guy from IBM
16:00.09mindjujucool
16:00.14Dezertoland he was very hard nosed about normal forms
16:00.22Dezertoland proper this .. and proper that
16:00.32Dezertolit was a very cool class
16:00.32mindjujui think eggyknap or thebigdog could prolly speak better to it
16:00.34Dezertolto be honest
16:01.00mindjujui've always wondered why pgsql and oracle don't offer auto-increment
16:01.11eggyknappgsql does, it just does it more sanely than MySQL
16:01.14mindjujueggyknap, dezertol's teacher said auto-increment is bad
16:01.19mgearyeggyknap: how so?
16:01.33mindjujuwas wondering why?
16:01.39eggyknapmgeary: see SERIAL and BIGSERIAL -- they're syntactic sugar for columns with default values based on sequences.
16:01.49mgearycool
16:02.25eggyknapmindjuju: there's a holy war in the database world (kinda like whether RISC is better than CISC) about primary keys that have meaning in the real world vs. those that don't. Most people will say your primary key should have something to do with the actual data.
16:02.52Dezertollike an "email address" or something
16:02.54eggyknap...but lots of times, you can't really make a unique key based on the data, so in those cases you have a synthetic key, based on a GUID, sequence, etc.
16:03.41eggyknapStrict relational theory claims that every row should have something unique about it, which might involve the combination of several columns (e.g. a multi-column primary key), and that unique something should be your primary key.
16:03.58mindjujuah!
16:04.00eggyknapIn some cases, though, it's easier just to make a generated column wholly unrelated to the data, and have it be your pkey
16:04.17unumthat's usually what I do
16:04.19eggyknapThat said, MySQL's auto_increment is bad :)
16:04.21thebigdogwhich is bad...its like crossing the streams
16:04.40Dezertolmost of the "big" projects I've used I generate my own base36 key
16:04.58unumI use the SERIAL datatype in creating my first column and call it tablename_index
16:05.39eggyknapOne of the major advantages of a meaningful pkey is that you have that much more data validation for "free". It's not as free as not having a pkey at all, but that often runs you into other troubles down the road.
16:06.08unumI have so few cases where I can use a meaningful pkey
16:06.12jsmithunum: Which is really just building an underlying sequence for you, and referencing that
16:06.18unumI have some situations where it seems like should be able to...
16:06.37eggyknapAnother advantage is that a pkey corresponds to an index, generally, and you want your indexes created on meaningful data.
16:06.57eggyknapheh
16:07.01eggyknapI need to finish my slides :)
16:07.45mindjujudo you need anything besides a projector?  white board or anything?
16:08.10unumin my projects table I have a jobnumber column and a year column  I have a unique restraint setting that requires the combination of those two columns to be unique
16:08.16eggyknapdepends on how many questions y'all come up with. Whiteboards are nice for ad hoc questions. Barring that, though, a projector is prolly fine.
16:08.52eggyknapWow... my 2 year old
16:09.16eggyknap...'s diaper just slid off. He said, "Daddy! My diaper just went 'Wheee... dunk.'!!"
16:09.32mgearywhee
16:09.32mindjuju:)
16:09.35eggyknapFortunately, its contents were non-messy.
16:09.36_carmonyeggyknap, then lets say I want to use a GUID in MySQL, what datatype should I use for fast comparisons on joins and stuff?
16:09.36Dezertolunfortunatly.. in my current project I could have 100+ rows with nearly "exact" data..
16:09.37unumbut I have to handle the projects prior to them getting assigned a job number, so I can't use that as my pkey
16:09.43mindjujukids... always good for a laugh
16:09.44Dezertoland so.. unique ids is my only option
16:09.55mgearyanyone have any insights into CentOS 5 versus Fedora 8?
16:10.14mindjujuhey guys, let's leave some questions for tomorrow!
16:10.25jsmithFedora 8 is already close to being end-of-life'd
16:10.26eggyknap_carmony: I'm not a MySQL guy, but GUIDs are probably much slower than something like auto_increment, because they probably require string comparisons (vs. numeric)
16:10.42mgearyjsmith: so CentOS 5
16:10.46jsmithmgeary: For a server, you want something like CentOS that's going to have 7 years of updates
16:10.56mgearysounds good
16:11.03mgearyCentOS is like RHEL, right?
16:11.05Dezertolhowever.. any uniq key like an email would also be a string
16:11.05jsmithmgeary: Fedora only has 13 months, as it's aimed more towards a desktop environment where you're going to upgrade
16:11.12jsmithmgeary: 100% binary compatible with RHEL
16:11.17mgearygotcha
16:11.21mgearygood deal, thanks
16:11.29Dezertolalso.. heard that if you want to do fast index searchs on emals
16:11.30jsmithmgeary: It's RHEL minus the red hat trademarks and images plus a few minor things
16:11.36Dezertolto create a second filed with the email backwards
16:11.40mgearyworks for me
16:11.56Dezertolso that you can do like "moc.oohay%"
16:12.16Dezertolas the index's work better in somet% then %something
16:12.27eggyknapDezertol: that's hackish, but works, if you're doing suffix searches like that.
16:12.47eggyknap(then again, lots of the things people have to do to get decent query performance can be hackish ;)
16:12.57Dezertollol
16:17.48eggyknap:)
16:18.09Dezertollol
16:18.15Dezertoli'm using http://search.cpan.org/~mwx/Data-Uniqid-0.11/Uniqid.pm
16:18.19Dezertolfor my id's
16:18.29Dezertoland i'm getting duplicate key violations
16:18.29Dezertollol
16:18.56Dezertol<PROTECTED>
16:19.09utahconBase62?
16:19.10utahconodd
16:19.32Dezertol3 import scripts runing at the same time pushing over 1mil each
16:19.40Dezertolon a very fast server
16:19.41Dezertollol
16:21.01Dezertoldon't have any problems till I kick the 3rd script
16:21.07macnewboldDezertol: maybe I missed this, but why aren't you doing it with auto_increment in the database?
16:21.19Dezertolcuz auto_increment is bad
16:21.20Dezertollol
16:21.24macnewbold?
16:21.26thebigdogDezertol, if you are using that script you might want to use luniqid() instead
16:21.31mindjujumacnewbold, his teacher said it was bad
16:21.34mindjujuwho was a guy from IBM
16:21.39macnewboldand?
16:21.52mindjujuand eggyknap explained that in normalization, you try to have all the data in the table relate
16:21.58macnewboldhere's case and point of why it isn't bad: you wouldn't be getting duplicate key violations :-D
16:21.59mindjujuusing an auto-increment doesn't relate to the data
16:22.10thebigdogDezertol, the issue with generating those is that you are bound to hit dups, you would have to keep track of them and then issue the command again on dups
16:22.11Dezertolthe biggest reason for me.. is that this records set has merging possibliities
16:22.22Dezertoland it's not worth fighting "auto_incrmeent"
16:22.25mindjujubut, eggyknap said it can be hard to find unique related data
16:22.30macnewboldmerging possibilities?
16:22.45*** join/#uphpu itimber (n=itimber@c-98-202-102-189.hsd1.ut.comcast.net)
16:22.45*** mode/#UPHPU [+v itimber] by ChanServ
16:22.59Dezertolmore then one source
16:23.04Dezertolcreating the data
16:23.09Dezertoland merging them together
16:23.19DezertolI'd end with two or more of the number 2 record
16:23.22macnewboldright, if they're all inserting into the same database, then autoincrement works great
16:23.43macnewboldif you're inserting them into separate tables/databases,
16:23.47macnewboldthen you'd need to renumber
16:24.12macnewboldeven with your uniqid or GUIDs or anything else, you'd just be playing a statistical game trying to reduce the likelihood of collisions when you merge
16:24.45macnewboldunless all the tables you start with have a unique identifier and the unique keys factor that in such that you don't get overlap from your data
16:24.56Dezertolnormally.. I run a custom generated base32 number
16:25.01macnewboldrandom?
16:25.05Dezertolbut.. in this case.. the table lock on where I keep track
16:25.10Dezertolis causing to much slowdown
16:25.21Dezertolso.. I opted for a uniquid
16:25.27macnewboldwell, you can have it done fast, or you can have it done correctly, but sometimes you can't have both :-D
16:25.29Dezertolbut now it goes so fast..
16:25.33Dezertolthat I'm getting dups
16:25.34Dezertollol
16:25.46macnewboldthen your unique ids aren't really unique
16:26.02macnewboldit's called a race condition, and locking will solve it
16:27.19Dezertolbasicaly.. my teacher who is/was the lead oracle dba for IBM
16:27.30Dezertolsaid never to use them
16:27.41Dezertoland every big project I've ever worked on
16:27.44Dezertoldoesn't use them either
16:28.42thebigdogDezertol, but your issue is not so much with the db but what you are trying to accomplish with the data
16:28.45macnewboldDezertol: if they were always the wrong answer, they wouldn't be in the database
16:28.50DezertolI have a little function I wrote that spins a base36 value that I've used for my keys (stole the idea from my last big project)
16:28.59macnewboldi.e. mysql wouldn't support them if they're not the right answer at least sometimes
16:29.12macnewboldin my experience, _most_ of the time they _are_ the best solution
16:30.05macnewboldin really large scale stuff there may be some drawbacks, but it is rare to have a problem that really must be of that scale
16:30.13utahcons/agress/agrees
16:30.26macnewboldyou can agress all you want, utahcon - I don't object
16:30.35Dezertollol
16:30.41mindjujuboth ids aren't related to the rest of the data
16:30.56Dezertolwell this database is already pushing over 30 million contact records
16:30.57mindjujuand one is more effective on byte level, and on a uniqueness level
16:31.04Dezertolit's a very large Lead database
16:31.09Dezertolfor an advertising firm
16:31.44Dezertollol
16:31.59Dezertolon the up side.. I can filter my self out ;-)
16:32.25utahconhmm
16:32.34utahconI think I need to update my bribe script.
16:32.38utahconnothing isn't of much value
16:32.39Dezertollol
16:33.26Dezertolmost of the records I use  $tmp = base_convert($do_key->KeyValue,36,10); $tmp++ $bump = strtoupper(base_convert($tmp,10,36));
16:33.48Dezertolbut because the spin on contacts table is so fast..
16:33.53DezertolI get table locks
16:34.01Dezertolon the key table where I keep track of what number im' on
16:34.18macnewboldso where are these numbers coming from? random? some hash on the data?
16:34.28macnewboldapparently they're running into conflicts because they're not sufficiently unique
16:34.28Dezertolthey increment but..
16:34.33DezertolI prepend a "site key"
16:34.42Dezertolfor which database generated the origonal record
16:34.43macnewboldif your site keys are unique,
16:34.48Dezertolcorrect
16:34.51macnewboldand you have a unique key for each record within each site
16:34.58Dezertolcorrect
16:35.00macnewboldsay like an auto_increment field for example,
16:35.12macnewboldthen a globally unique key is made by concatenating both keys
16:35.16macnewboldthen you can merge freely
16:35.18Dezertolcan't prepend the site key to the auto_increment
16:35.18macnewboldno conflicts
16:35.27Dezertolo..
16:35.30Dezertolya.. could..
16:35.31macnewboldDezertol: yes you can, it's easy
16:36.11Dezertolnot sure about splitting it up into two fields
16:36.44macnewboldif you don't like two fields, you can make them into one
16:36.55Dezertolbut then it's not "auto_increment"
16:37.03macnewboldyou only need the combined globally unique key when you try to merge
16:37.03Dezertoland by using base36
16:37.22macnewboldwhat does base36, base32, base62, or anything else have to do with it?
16:37.45macnewboldwhy not just use an integer? computers are really fast at dealing with integers, even really big ones.
16:37.45*** join/#uphpu cycokilr (n=cycokilr@216.112.116.6.ptr.us.xo.net)
16:37.50macnewboldthey store them really densely too
16:37.58Dezertolusing base36 I can have 01A as an id
16:38.27macnewboldand using an integer, you can have 1225 as an id
16:38.45Dezertolonly i'm at 40,000,000 at this point
16:39.06macnewbold(sidenote: generally, a good rule of thumb is to use the simplest solution that works correctly)
16:39.20macnewboldso what's wrong with having 40000000 as an ID? what makes 01A better?
16:39.27mindjujurube goldberg ala sql
16:39.32macnewboldif you say it prints out in less space, I'm going to cry
16:39.38Dezertollol
16:39.41Dezertolno that's not it
16:39.53Dezertoltho.. I thinking back.. I wish I owuld have
16:39.54Dezertollol
16:40.21Dezertolbasically.. all the big projects I've worked on don't use them and my dba teacher told me never to use them
16:40.27Dezertolso.. I've jsut never used them
16:40.28Dezertollol
16:40.58macnewboldintegers? or auto increment?
16:41.04Dezertolauto increments
16:41.17macnewboldnever believe a person who has never used auto_increment about how useless they are...
16:41.36macnewboldanyway, so why won't an integer ID work for your data?
16:41.37dataw0lfDoesn't really matter what numeral system you use.  Probably best to use one that you're most familiar with though...
16:42.02macnewboldyeah, my brain works best with base 10 or base 2. base 37.5 not so much
16:42.11Dezertollol
16:43.00mgearyunless you have 37.5 fingers, like mindjuju
16:43.07mgearyin which case, it's cake
16:43.14macnewbold~cake
16:43.15jbotthe cake is a lie!
16:43.16mindjujui love cake
16:43.21macnewbold~browies
16:43.25Dezertollol
16:43.25mindjujuespecially the song, going the distance
16:43.27macnewbold~brownies
16:43.28jbotMaaaaaaaaaaaan i've got the munchies!
16:43.41mgeary~twizzlers
16:44.49mindjujuhere is the song - http://www.youtube.com/watch?v=__PU5CVSegg
16:45.48mindjujueven thebigdog loves that song
16:47.00thebigdogwell yeah
16:47.00Dezertolmacnebold just an fyi.. Oracle doesn't have an auto_increment ability
16:47.00mindjujucourse the vid has a Talking Heads -- Road to nowhere feel
16:47.29mindjujuDezertol - i heard they have something called sequentials
16:47.34mindjujuthough i really don't know what that means
16:47.41mindjujuseems like josephscott mentioned it once
16:47.47mindjujuor maybe it ws thebigdog
16:48.01Dezertolhttp://hocuspokus.net/2007/10/auto-increment-ids-in-oracle
16:48.05Dezertolyou have to create it..
16:48.12josephscottin DBs there are two general approaches to that feature, auto-inc fields and sequences
16:48.14eggyknapLooks like $me has been misunderstood, earlier. auto_increment is probably faster and more reliable than <insert your UID generator here> if you need a synthetic primary key on some single table. Prefer primary keys that involve your actual data to synthetic keys, and use some sequential integer keys for synthetic keys in preference to some UID if you're staying in one table.
16:48.58josephscottDBs that use auto-inc are ms access, mysql, ms sql server
16:49.14josephscottDBs that use sequences are DB2, Oracle, Sybase, PostgreSQL, etc
16:50.09josephscottFrom a scale issue I believe that conceptually sequences will perform much better than auto-inc fields
16:50.47thebigdogseq are data objects that are really nice to work with
16:50.53josephscottat a high level the user mostly treats them the same, but their order of use is different
16:50.56eggyknapAn advantage of using a sequence is you can use that same sequence on several different tables simultaneously
16:51.17josephscottfor auto-inc you get the new ID after the insert
16:51.21eggyknapNote that in PostgreSQL, though, sequences aren't transactional, so although you won't get duplicate values from a sequence, there might be gaps in your sequence.
16:51.22thebigdogand you can write logic around them too which is sometimes beneficial
16:51.30josephscottfor sequences you get the new ID first and then use it as part of the insert
16:52.05thebigdoglike eggyknap they are not transactional...so if your insert fails your seq has incremented
16:52.15josephscotteggyknap: I think most implementations of both auto-inc and sequences warn that there is no guarantee that there won't be gaps
16:52.59eggyknapNote that for sequences you can (in pgsql anyway) set a field's default to the next value of that sequence, so you don't have to grab a sequence value first and then insert a row in a second query.
16:53.28josephscottyep
16:54.03josephscottI think that's probably true of all DBs that use sequences, because you can implement that feature as a trigger which I believe most (all?) DBs that support sequences support (triggers that is)
16:56.51*** join/#uphpu cycokilr (n=cycokilr@216.112.116.6.ptr.us.xo.net)
16:58.45herloInteresting Idea I'm thinking about trying to do once a month: http://www.workatjelly.com/
16:59.53*** join/#uphpu beandog (n=steve@gentoo/developer/beandog)
16:59.53*** mode/#UPHPU [+v beandog] by ChanServ
17:00.05macnewboldregardless of whether you use auto_increment or sequences as the mechanism, the principle is the same: if you want unique IDs, set up a pattern that prevents duplication, like take the last ID and add 1 to get the new ID, with some locking to make sure two people don't grab the same next ID
17:00.21beandoguuid++
17:00.21macnewbold~s/people/processes or threads/
17:00.35*** join/#uphpu _psychic__ (n=_psychic@173.8.91.177)
17:00.48Dezertolya.. I was doing that but the locking and unlocking was taking to long
17:00.48Dezertollol
17:06.22josephscottbeandog: uuid is good if you need maximum uniqueness, but in some cases when using MySQL it can be very inefficient
17:06.40beandogjosephscott, agreed.
17:06.41josephscottDezertol: came into this a bit late, what are you storing this in?  MySQL?
17:06.52Dezertolya
17:07.01Dezertolit's a lead database for an ad firm
17:07.04josephscottwhat table type? myisam?
17:07.07Dezertol40 milion strong
17:07.09Dezertolmyisam
17:07.15josephscottwell there's your problem :-)
17:07.20Dezertollol
17:07.39josephscottany particular reason you aren't using innodb?
17:07.52Dezertolthe data is almost 100% insert
17:07.59Dezertoland myisam is better tuned for that
17:08.28Dezertolwell.. insert and select
17:08.31Dezertolno updates
17:08.33Dezertolno deletes
17:09.20Dezertolthe issue is that at any given time there could be 10+ csv's importing at the same time
17:09.23josephscottif you are having locking problems (which is easy to do with myisam) you can likely scale better with innodb
17:10.19josephscotthistorically innodb does have some particular use case problems, where the only realistic option (under MySQL) was to use myisam, but the number of cases where that is true is shrinking
17:10.59Dezertolwell I know the locking is different
17:11.04josephscottDezertol: myisam is having to do a full table lock on those inserts
17:11.09Dezertolbut honeslty.. i've not spent much time with innodb
17:11.15josephscottwhich innodb won't (in most cases) have to do
17:11.43DezertolI only need to lock the table/row where the key is stored
17:11.50josephscottalso, at 40mil rows you probably start to see other pains
17:11.51DezertolI bounce it on each pass and store the new value
17:11.55Dezertolbase36
17:12.13Dezertolit works fine.. but it's slow to lock and unlock each time
17:12.30Dezertolwith several keys per insert being created
17:12.40josephscottyep, innodb does better with locking (MVCC for the win!)
17:13.19Dezertol?mvcc... Military Vehicle Collectiors of California?
17:13.46josephscottMultiversion concurrency control
17:13.50josephscotthttp://en.wikipedia.org/wiki/Multiversion_concurrency_control
17:15.12Dezertolintresting..
17:15.43josephscottwhen's the last time you backed up the table with 40mil rows in it?
17:16.04Dezertolit backs up nightly
17:16.21Dezertoland the data that is imported is from csv's
17:16.23mgearybet that takes a while
17:16.25Dezertolso if it fails..
17:16.31DezertolI can just re-load the csv's for the da
17:16.33Dezertolday
17:16.38Dezertolit's not bad actually
17:16.49Dezertolthe server.. is.. well.. custome built for this
17:17.01Dezertolit's a quick box
17:17.21Dezertolif it was slower..
17:17.24eggyknapOuch. pgsql + PITR FTW there.
17:17.27DezertolI might not have this problem
17:17.28Dezertollol
17:19.06josephscottwhat's interesting about postgresql's storage techniques is that the amount of disk needed to store those 40mil rows will likely be less than what mysql is using
17:19.33mgearywow, it's nice out there
17:19.40eggyknapjosephscott: why's that, out of curiosity?
17:19.56utahconmgeary: stop that!
17:20.04Dezertollol
17:20.12mgearyutahcon: seriously
17:20.25utahconI hate working a day job when the weather is all nice
17:20.28DezertolI actually called MySQL and talked to one of ther guys to make sure it would be ok for this project
17:20.39josephscotteggyknap: couple of reasons usually, one is the compression techniques pgsql uses
17:20.57josephscottthe other is how mysql will often duplicate primary key indexes
17:20.57eggyknaptoast is nice, yes :)
17:21.05eggyknapIndeed? Ouch.
17:21.37josephscottthe key duplication thing helps on lookups, but can increase disk usage a fair bit
17:22.33Dezertolonly useing 20g of a 1.1T on that server
17:22.44Dezertollol
17:23.02mgearyDezertol: how long do you keep your nightly backups?
17:23.26*** join/#uphpu _psychic_ (n=_psychic@173.8.91.177)
17:23.26*** mode/#UPHPU [+v _psychic_] by ChanServ
17:23.31josephscottDezertol: a few years old, but interesting case: http://feedlounge.com/blog/2005/11/20/switched-to-postgresql/
17:23.51DezertolI sync the weeks's end off the server
17:23.54Dezertolbut that's bout it
17:24.06eggyknapjosephscott: DO you know, right off, anywhere I can read about that one? It's a MySQL-ism I'd like to get my head around.
17:27.38josephscotteggyknap: I think that feedlounge post has comments that talk about it
17:27.54josephscotteggyknap: "InnoDB stores the uncompressed primary key in all secondary index records."
17:28.02eggyknapthx.
17:28.03josephscotthttp://feedlounge.com/blog/2005/11/20/switched-to-postgresql/#comment-406
17:29.21eggyknapWow...that's really ... interesting. I can see how that might make lookups much faster, in some cases. But wow.
17:30.50josephscottgoes back to our conversation the other day, figure out what your needs/challenges are and build accordingly
17:31.02Dezertollol speaking of db.. just got this sessions fortune: ->
17:31.05DezertolHas everyone noticed that all the letters of the word "database" are typed with the left hand?  Now the layout of the QWERTYUIOP typewriter keyboard was designed, among other things, to facilitate the even use of both hands. It follows, therefore, that writing about databases is not only unnatural, but a lot harder than it appears.
17:31.27josephscotthaha
17:31.38jsmithGood one!
17:39.17*** join/#uphpu supernovia (n=Velda@c-67-164-250-20.hsd1.ut.comcast.net)
17:45.35mindjujuso if i'm going to do software RAID in ubuntu, i should use Linux's Multiple Device (md) raid system, optionally with the Logical Volume  Manager (LVM).
17:55.51jsmithmindjuju: Aye...
17:57.11mindjujuthx jsmith
18:17.25Dezertol?.. is software raid even worth doing?
18:20.41mindjujuit is when there isn't a hardware raid option
18:20.47mindjujuand you want ome protection, methinks
18:21.18Dezertolo..
18:21.23Dezertollol ya mirrored options
18:21.45Dezertolfor whatever reason my brain was like... if it's not got a raid card.. I just assumed you'd strip it for speed
18:21.47Dezertollol
18:22.08Dezertolbut ya mirroring works
18:22.10josephscottmuwahahah http://www.microsoft.com/web/gallery/
18:23.23Dezertolo.. dear.. "DotNetNuke"
18:23.28Dezertollol that about sumes it up doesnt' it
18:23.29Dezertollol
18:24.28SunSparcAll nuke derivatives need to be nuked.
18:24.43mindjujuCan't we all get along? From ASP.NET to PHP, IIS 7.0 provides a powerful and flexible Web server for  the world’s most popular Web applications. ASP.NET hosting is even more simple  and powerful, with deep integration directly in the request processing pipeline,  configuration system, and management console.
18:25.03mindjujufrom here - http://www.microsoft.com/web/platform/server.aspx
18:25.38mgearyand, best of all, you get to pay us every month, just for letting you use our wonderful software, which you also paid for when you first got it
18:26.13DezertolI recently turned down a web job they wanted me to do some custom work on there site.. lol it was php3's version of phpnuke
18:26.13Dezertollol
18:26.17DezertolI told them no
18:26.17Dezertollol
18:26.37Dezertolstill had the php3 file names and everything
18:26.39Dezertollol
18:26.45mgearygah
18:26.55tierramindjuju: http://www.howinthetech.com/quick-and-dirty-linux-software-raid5/
18:27.11tierraheh
18:27.38*** join/#uphpu mindjuju (n=mindjuju@76.8.222.130)
18:27.38*** mode/#UPHPU [+o mindjuju] by ChanServ
18:28.14tierramindjuju: http://www.howinthetech.com/quick-and-dirty-linux-software-raid5/ - just switch out the --level=raid1 for mirrored
18:29.37tierrasoftware RAID in linux has become pretty stable and easy to setup these days
18:32.46DezertolIIS Media Services, an integrated HTTP-based media delivery platform, delivers true HD (720p+) streaming and provides real-time logging to measure media investments. By offering a complete media delivery platform together with a traditional Web server, rich dynamic Web sites can now be managed and administered from a single Web platform
18:33.31Dezertolwouldn't it be the "content at 720p+" that would be delivered?
18:33.37mindjujui'm streaming right now with winders media stream 9
18:46.47itimberquestion: what's the best way to check that a MySQL datetime field is > now
18:46.57itimber`datetime` > NOW()
18:46.59itimber??
18:48.09itimberor do I need to do something like this: YEAR(`datetime`) > 2009 AND MONTH(`datetime`) > 03 AND etc....
18:48.10fungusitimber: yup
18:48.17itimberTY
18:48.24fungusa simple > works fine with dates
18:57.18*** join/#uphpu thebigdog (n=thebigdo@207.135.132.13)
18:57.19*** mode/#UPHPU [+o thebigdog] by ChanServ
18:59.08*** join/#uphpu DexterTh1Dragon (n=admin@71-35-207-73.slkc.qwest.net)
19:46.00thebigdogbeandog you around?
19:58.32*** join/#uphpu carmony (n=justinca@c-67-164-249-224.hsd1.ut.comcast.net)
19:58.32*** mode/#UPHPU [+v carmony] by ChanServ
20:01.24thebigdogi could go for some great chinese food right now
20:01.32thebigdogbut i have too much work
20:01.56mindjujui can't think of a Chinese restaurant that doesn't deliver
20:02.03mindjujulike i think it is required by law or something
20:04.43thebigdogi think it is required by law
20:04.52supernoviaI need to review the not to do list
20:04.57thebigdogthat is like form b
20:05.53supernoviaand I'll add to it, "fantasize about a lunch at Chine Isle"
20:06.44thebigdogi think mindjuju is about ready to sale away with cpt. morgan
20:06.54thebigdogs/sale/sail
20:07.30*** join/#uphpu Utah_Dave (n=David@72-166-51-162.dia.static.qwest.net)
20:07.30*** mode/#UPHPU [+o Utah_Dave] by ChanServ
20:07.39supernoviahi Utah_Dave
20:07.54mindjujuyeah, thebigdog, i think you've been "saling" away with Capt' Morgan this morning
20:08.02mindjujulittle rum and coke maybe?
20:08.18thebigdogmindjuju, the problem is - im out of coke
20:08.40mindjujui've never tried any of it, but i hear rum and pepsi is bad
20:08.48mindjujulike don't try it bad
20:09.36Utah_Davehey, supernovia
20:09.57Utah_DaveI had rum and coke on my last flight back from D.C.
20:09.58supernoviaI still need to pick up some old boxes from you :-D
20:10.14Utah_DaveI asked for a coke, and the stewardess brought me a rum and coke
20:10.18supernoviaNice
20:10.26Utah_DaveI took one sip.
20:10.34thebigdogUtah_Dave, that is so funny
20:10.45Utah_DaveI don't think I hold my alcohol very well   :(
20:10.47mindjujuwas you in first class?
20:10.50thebigdogman if you had an alergic reaction i see $$$$
20:10.52Utah_Daveyeah, first class
20:11.09supernoviahehe
20:11.11mindjujui walked through first class once
20:11.20Utah_Davethe stewardess was freaking out. She kept saying, "I'm so sorry!! I'm so sorry! I didn't mean to force that on you!!!"
20:11.21mgeary~lart mindjuju for saying "was you in"
20:11.21jbotslaps mindjuju around with a large trout for saying "was you in"
20:11.26supernoviaI've walked through 1st class lots o times. every time I fly even
20:11.38mindjujumgeary, got to axe you a question
20:11.42mindjujuwhy you be hating?
20:11.43Utah_DaveI got upgraded to first class both ways on that trip
20:11.45mgearykk
20:11.45mindjujuor is that hatting?
20:11.58supernoviano dude
20:12.01supernoviait's h8ting
20:12.07Utah_Davesat 2 seats in front of Sen. Bennet on the way out and next to the Treasurer of Huntsman corp on the way back
20:12.07supernoviaor h8ing maybe ..
20:12.20mindjujuthat's awesome utah_dave
20:12.32mindjujuhoping and hopping always make me do a double take too
20:13.13supernoviadon't think I"ll be flying 1st class anytime soon.
20:14.16mindjuju9 years
20:14.32supernoviawowsas.
20:14.38supernoviahate to fly?
20:14.47mindjujui love flying
20:14.50mindjujui love traveling
20:14.51*** mode/#UPHPU [+v DexterTheDragon] by ChanServ
20:15.00mindjujuflying || traveling == $$
20:15.29supernoviaor jobs that make you fly and travel...
20:15.44mindjujuoh wait, you're right, i flew to cali on bidness
20:15.50mindjujubut that was last year
20:16.31supernoviaI got to go to hollywood last month
20:17.06mindjujuwow - He said he and his economic team have begun discussions with leading  congressional players to fast-track legislation creating another regulatory  entity to govern the dissolution of large financial institutions such as AIG, which if merely  liquidated could reverberate disastrously through the financial system.
20:17.53mindjujuall talk about additional bureaucracy aside, looks like maybe the worst hasn't hit yet
20:18.06mindjujuhttp://news.yahoo.com/s/ap/20090318/ap_on_go_pr_wh/obama_aig
20:20.26mgearyget rid of the whole "too big to fail" thing
20:21.00mindjujulol
20:21.14mgearyi've got yer bonus right here
20:22.00utahconreally? woot!
20:22.33mgeary*POW*
20:22.47supernoviaooh  yeah, I've had that bonus before.  The joy is all in the anticipation.
20:23.11mgearyi heard today that the bonuses are between $1k and $6mil
20:23.36utahconthat is a lot of execs
20:24.03utahcon"11 people who got these bonuses are gone!"
20:25.18itimberI'm out guys.... see ya'll
20:27.14mindjujuanybody see a deja vu moment with this "And, you know, he is doing so with intelligence and diligence. Nobody's working  harder than this guy," Obama said.
20:27.32mindjujuand this
20:27.33mindjujuBrownie, you're doing a heck of a job,
20:27.58mgearyheh
20:28.38mindjujuyeah true that
20:29.01mindjujuugh, 30 minutes for lunch goes much too fast
20:29.10im_busier_than_jbah
20:29.23thebigdoghttp://www.hulu.com/watch/28343/dr-horribles-sing-along-blog
20:29.32SunSparcMaybe we should all ping flood jsmith-busy...
20:29.40busierthanjsmith~pester jsmith-busy
20:29.40jbotjsmith-busy: Are we there yet? .. Are we there yet? .. Are we there yet?  see http://beverlys.net/LJ/BuggingYou.swf
20:29.46mindjujuthebigdog - that's so 2008
20:30.03thebigdogoh i know but man it is still hilarious
20:30.35jsmith-busyYou guys rock!
20:30.43jsmith-busy(Oh, and the gals too)
20:30.59thebigdogi call my buddy bait and switch all the time
20:31.02SunSparcjsmith-busy: Why, because we are busier than you?
20:31.17SunSparc;)
20:31.35jsmith-busySunSparc: No, because I was grumpy that I had to ignore IRC and get more work done, but I can't possibly ignore the window if it's blinking!
20:32.01thebigdogi need to start quoting that more
20:32.29SunSparchehe
20:32.35SunSparcjsmith-busy: Glad we could help.
20:45.59supernoviaoh and sorry for making it blink.
20:59.40*** join/#uphpu beandog (n=steve@gentoo/developer/beandog)
20:59.40*** mode/#UPHPU [+v beandog] by ChanServ
20:59.53supernoviahi beandog
21:00.03beandogsupernovia, buenos
21:00.10supernoviaoh and hi SunSparc too :) thanks for the tweet reply
21:00.22SunSparc:)
21:00.28supernoviamas o meno beandog :-)
21:00.32supernoviabut to you too :o)
21:00.48beandogma o meno
21:00.57supernoviaah, ma o meno. Thanks.
21:01.14supernoviais it really without the S? I always thought that was just how it was pronounced by some people
21:01.30beandogthey just pronounce it without the s
21:01.42supernoviain which countries or is that everywhere?
21:01.53beandogargentina, at least
21:02.11jsmith-busyargentina and chile are apt to leave off the 's'
21:02.12supernoviahah, good to know.
21:02.21beandogque loco
21:02.21supernoviamy brother in law went to argentine
21:02.26supernoviaargentina and he always makes fun
21:02.33supernoviaof the way supernovio leaves off s's
21:02.35jsmith-busyand chile is very apt to leave out a "d", like in "tostao" instead of "tostado"
21:02.43supernoviayeah
21:02.46*** join/#uphpu mgeary (n=mgeary@166.70.44.49)
21:02.46*** mode/#UPHPU [+v mgeary] by ChanServ
21:02.49unumI thought the further north you went the more letters they left out
21:02.51beandoghola, mgeary
21:02.53jsmith-busyparaguayans just slaughter the language completely
21:03.02unumby the time you got to the caribean they only prounce the first letter of the word
21:03.18supernoviaooh yeah.
21:03.22mgearywill someone please think of the languages!?
21:03.23unumparaguayans use more native words than spanish
21:03.35beandogspanish needs more slang
21:04.21jsmith-busyunum: And they slaughter spanish words when guarani doesn't have a corresponding word... like kabaju for caballo
21:06.43mgeary~granola supernovia
21:06.54supernoviagar.  found myself trying to tab out of that ..
21:06.56beandogick
21:07.08beandoggranola is taking a bunch of stuff that tastes good alone, and making it taste meh together.
21:07.30supernoviaoh goodness. you haven't had that vanilla almond granola from sunflower market I take it
21:07.38beandogwhat's sunflower market?
21:07.51beandogis that where you hippies shop?
21:08.27mgeary~birkenstocks supernovia
21:08.30beandogmeh, closts is in murray
21:08.32thebigdogbeandog, you there?
21:08.34beandogclosest
21:08.36beandogthebigdog, yah
21:08.47beandogI'm addicted to CLIF bars.
21:08.49supernoviathere's one in Orem
21:08.55beandogI'm in Salt Lage
21:08.57beandoger
21:08.58beandogLake.
21:09.03beandoghow did I get a g ...
21:09.07supernoviaso whenever I go to provo I pick up some fresh fruits and veggies :-)
21:09.13supernoviagood q!
21:09.19beandogwhenever I go to provo, I pick up some hawties.
21:09.40supernoviacause provo's just full to brimming with those eh?
21:09.47beandogwell, tehre arent any up here
21:09.50beandogheh
21:10.00supernoviaheh
21:10.18thebigdogi love argentina
21:10.28thebigdoghawties with the bad teeth
21:11.29supernoviaanyway mgeary, I'm not as granola as you'd think, but I do wear sandals as long as there's no more than 1 inch of snow. and I am all over veggies and stuff ;-)
21:11.40mgeary~cutoffs supernovia
21:11.46supernoviabut mainly because I'm constantly on a diet :-p
21:11.56supernoviaand veggies really are good :-)
21:12.05beandogwouldnt granola be really fattening?
21:12.07jsmith-busythebigdog: and Guarana!
21:12.12supernoviait is actually
21:12.34thebigdoghehe
21:12.41mindjujuisn't guarana another name for bat poop?
21:12.47supernoviaguarana is so good!
21:12.54mindjujuyuck!
21:13.03supernoviabut yeah it's fattening, gotta get some calories with the raw food thing ya know
21:13.12supernoviaGee, thanks.
21:13.43beandogmgeary, what about legpits
21:14.21supernoviaI hope you refer to the space behind one's knees..
21:15.05beandoguh, yah.
21:15.06beandogo.O
21:15.16supernoviao.O is right!
21:15.33beandogI didn't say ... oh, well, there is no good word for that is there
21:15.49supernoviano appropriate word anyway :-p
21:15.55beandoglol
21:16.00beandoghow about, mudflap
21:16.02beandog:)
21:17.36beandogyah, ah well
21:17.45supernoviaso uhh...
21:17.47beandogso
21:17.50supernoviayoga
21:17.55supernoviais good
21:18.01beandoganyone know a good address book manager in PHP
21:18.38supernoviahmm
21:18.49utahcondocument locker?
21:18.56beandogutahcon, yah
21:18.56beandoglike
21:18.59utahconlike... upload to folder, display download links?
21:19.03beandogupload PDFs, docs, etc.
21:19.06beandogpassword protect em
21:19.08beandogkinda thang.
21:19.10utahconhmm
21:19.24utahconI wrote something like that a few months back... I have no idea what I did with it though
21:21.02beandognever written one before
21:21.15beandogit seems relatively simple to design, but nothing like getting some ideas
21:22.04utahconI went over the top and made it all OOP and what not.
21:22.25utahconI also added some options as to where the file would be stored, either on my server, or in S3 from Amazon.
21:25.42beandognever used S3
21:25.45beandogis that any good?
21:25.58beandogI think I remember seeing a fuse for that one.
21:26.06beandogsys-fs/s3fs
21:26.18beandoghttp://code.google.com/p/s3fs/
21:26.33beandogseems like it'd be slow to me, any network based drive.
21:26.40beandoger, over the internets.
21:27.00utahconI was using S3 to cut down on bandwidth costs for large files I was distributing,
21:27.05utahconFor that it was great.
21:27.15beandog"large files" ?
21:27.23beandog:)
21:27.29beandogutahcon, thats cool.
21:27.32utahconall (mostly) legit :d
21:28.12utahconI also had my personal music storage on there, for a short while, and using the FUSE you mentioned was able to have all my musics everywhere I went.
21:28.43beandoghmm
21:28.46beandognow that would be a good idea.
21:29.07utahconI occasionally move a large file from location to location that way, but not as often anymore.
21:29.37utahconNice thing about S3 is you only pay for what you are using. So if you upload 10GB of files one month and remove them all before your next billing cycle, then you pay nothing in month 2.
21:29.41beandogoh yah, humm
21:29.47beandoghow much does it cost, anyway
21:30.07utahconum...
21:30.37utahconhttp://aws.amazon.com/s3/
21:30.47utahconstarts at $0.15/GB
21:30.52utahconplus transfer
21:30.54utahconand requests.
21:31.06beandogfreek
21:31.08beandogthats cheepy.
21:31.12utahconit can be
21:31.32utahconI think there are other options out there too, but I like Amazon. They have always been good to me.
21:35.52utahconI would probably hit uber if they gave me a kindle.
21:35.58utahconstill then I am uber-1
21:36.20utahcon/s/still/till
21:36.25utahcon/s/still/till/
21:36.36utahcondoes the bot not catch corrections?
21:36.46utahcon~lart fungus
21:36.46jbotbeats fungus senseless with a 50lb Unix manual
21:37.10beandogI think you just use s/
21:37.13fungusno leading slash on s/search/replace/
21:37.22utahconoh
21:37.26utahconok
21:37.30fungus~whaleslap utahcon
21:37.30jbotACTION beats utahcon upside and over the head with a freakishly huge killer whale named Hugh
21:37.47utahconwho knows their simplexml?
21:37.58beandogI know simplexml_load_string
21:37.59beandogand thats about it.
21:38.03utahcondoh
21:38.11thebigdogwhat do you need help with Utah_Dave
21:38.14utahconI need to access the attributes, but I can't seem to recall how the hell to do it.
21:38.19thebigdogwhat do you need help with utahcon
21:38.54thebigdogattributes()
21:38.58utahconthanks!
21:39.09thebigdog$obj->attributes()
21:39.14thebigdoggives you a list of them
21:39.38thebigdogyou can probably use list to get them
21:39.59utahconthat is just odd syntax $file->attributes()->id
21:40.17thebigdogyeah
21:40.28thebigdogbut that is standard object notation
21:41.18thebigdogwhat does $file['id'] give you?
21:41.48thebigdoggetAttribute('id') might work too
21:42.01thebigdog$file->getAttribute('id')
21:43.05Utah_Davethebigdog: must have been someone else needing help
21:43.10Utah_DaveI don't remember asking
21:43.24thebigdogsorry Utah_Dave - tab completion
21:43.40jsmithSorry Dave... I can't let you do that
21:44.04Utah_Dave:)
21:54.26brastoisn't there a directive in sendmail to send all outbound mail to a specific host for processing?
21:54.36Dezertolya
21:54.40Dezertolit's the smarthost
21:54.51Dezertolsearch the config file for smart host
21:55.59Dezertolsmart host.. two words
21:57.24brastoDezertol: so uncomment this line in sendmail.cf: #O FallbackSmartHost=fall.back.host.net
21:57.34Dezertolit's not the fallback one
21:57.43Dezertolthere should be anohter
21:58.29brastoR$* < @ [ $+ ] > $*     $: $1 < @ [ $2 ] : $S > $3      Add smart host to path
21:58.41brastothat's the only other line with two words 'smart host'
21:59.01mgeary~ping mindjuju
21:59.02jbotpong mindjuju
21:59.22Dezertolhttp://www.cyberciti.biz/faq/configure-sendmail-as-a-smart-host/
21:59.23brastodid i overlook something Dezertol
21:59.25brasto?
21:59.34brastolooking at link...
22:02.39Dezertolon slack 12 the mc file is /usr/local/src/cdrom/source/n/sendmail/sendmail-slackware.mc
22:03.36Dezertolnot very big 42 lines.. most of which is comments
22:03.41brastomaybe I wasn't finding it since I was looking at sendmail.cf instead of sendmail.mc
22:03.55Dezertolya.. the cf is the "generated" file
22:04.18Dezertolwhich is why postfix exists.. he got sick of how sendmail does things
22:04.19Dezertollol
22:07.42brastoyea, I've worked more with sendmail, but when I've done work with postfix, I did find it easier to do things with.
22:08.21Dezertolya I stoped using sendmail 7ish years ago.. in favor of postfix
22:08.36Dezertolgrey listing is awsome
22:09.03Dezertolhttp://postgrey.schweikert.ch/
22:09.14mindjujumgeary, i have achieved "single user mode"!
22:09.25mindjujuforgot which one was command for a sec, so it no worky
22:09.30mgearysee pm
22:43.48beandogqmail++
22:44.11beandogIve never gotten *any* mail server to work as fast or as easy as qmail
22:48.23tierraditto, though postfix and exim aren't too hard to configure in Debian/Ubuntu
22:48.56tierrabut I do still run a qmail debian server
22:55.25thebigdogpostfix is the diggity bomb
22:56.26dataw0lffo shizzle
23:00.41unumexim just works when I install debian
23:00.47unumso I've always just used it
23:03.25orsonjyep, have to like "turn on email (y/n)"
23:08.18beandograwr
23:08.22beandogbestbuy's website is caca.
23:08.42brastohttp://bestbuy.com
23:08.44brastodoh
23:08.59beandogcaca!
23:10.35brastoi don't see anything wrong...
23:11.02beandogits just not very friendly
23:11.05beandogsearch is stinky
23:11.11beandogit doesnt remember your zip on find in store
23:11.11brastoah
23:11.13beandogetc.
23:11.36brastoah, they need training from someone in uphpu then
23:12.26brastowps or mgeary can show them how to do their site right.
23:12.53mgearyheh
23:14.14brastoi see their main problem.  They're using jsp
23:14.34brastothat's a bigger problem than not remembering zip codes for local searches...
23:14.36brastojk
23:15.12beandogneeds moar 2.0
23:15.24beandoglike dunder mifflin infinity
23:16.36macnewbold~rlart
23:16.39jbothurls dozens of incontinent, insomniac, hungry kittens with tiny little razor-sharp claws and a wide variety of contagious intestinal parasites at SunSparc
23:16.39macnewbold~yow
23:16.54beandoggood ol SunSparc
23:18.01SunSparcand on that note...
23:18.05SunSparcLaterz yall!

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.