IRC Logs for #crux Monday, 2007-10-22

*** MzOzD has quit IRC00:02
*** mwansa has joined #crux01:20
*** pitillo has joined #crux01:28
pitillogood morning01:28
*** mwansa_ has quit IRC01:39
*** mwansa has quit IRC01:40
*** pitillo has quit IRC01:46
*** pitillo has joined #crux01:46
Romsterhi01:48
*** onestep has joined #crux01:57
*** sepen has joined #crux02:03
*** Roomster has joined #crux02:52
*** onestep has quit IRC02:57
*** roliveira has quit IRC03:01
*** onestep has joined #crux03:01
*** Romster has quit IRC03:02
*** Roomster is now known as Romster03:02
*** RedShift has joined #crux03:14
*** namenlos has quit IRC03:18
*** namenlos has joined #crux03:21
namenloshi, enemy territory: quake wars full linux client has been released *jump*03:24
* tilman heard etq sucks03:24
tilmanetqw even03:24
tilmannice timing though, only 3 weeks late, right?03:25
namenlosyes, something like that...03:28
rxitilman: its pretty awesome03:46
namenlosi like it, too03:47
namenlosrxi: you are playing it on linux?03:48
*** onestep has quit IRC03:52
rxinah i played the demo on win3203:55
rxitempted to buy it tho03:55
namenlosah, ok, because i got some troubles launching it on linux...03:56
rxiid stuff is usually pretty painless03:57
namenlosdoom3 was...03:58
RedShifthey guys, you got experience with mod_ruid?04:00
RedShiftit seems too good to be true04:00
*** j^2 has quit IRC04:10
*** onestep has joined #crux04:26
*** j^2 has joined #crux04:26
*** mwansa has joined #crux04:31
*** onestep has quit IRC04:38
*** onestep has joined #crux04:39
*** onestep has joined #crux04:39
namenlosman, man, man, i was able to install and launch etqw, but in-game there aren't any fonts...05:00
*** minux_ has joined #crux05:01
pitillolinux sin efectos = nada?05:04
pitillosorry05:04
tilmannp05:05
tilman;)05:05
*** minux has quit IRC05:05
pitillobrbrbrbr welcome back to irssi and /window X :)05:08
*** treach has joined #crux05:32
*** onestep has joined #crux05:33
*** treach has quit IRC05:59
*** treach has joined #crux05:59
*** treach has quit IRC06:08
*** mike_k has joined #crux06:22
*** onestep has quit IRC06:42
*** onestep has joined #crux06:47
*** namenlos has quit IRC06:58
*** kieselsteini has joined #crux07:04
*** onestep has quit IRC07:17
*** kieselsteini has quit IRC07:21
*** kieselsteini has joined #crux07:23
*** onestep has joined #crux07:25
*** kieselsteini has quit IRC07:46
*** mxq has joined #crux07:50
*** minux_ has quit IRC07:59
*** namenlos has joined #crux07:59
*** jdolan_ has joined #crux07:59
*** jaeger has joined #crux09:10
*** ChanServ sets mode: +o jaeger09:10
*** mxq has quit IRC09:13
*** triel has quit IRC09:43
*** triel has joined #crux09:50
*** boe has joined #crux10:08
*** onestep has quit IRC10:16
*** onestep has joined #crux10:17
*** boe has quit IRC10:39
*** destruct_ has joined #Crux10:48
*** destruct has quit IRC11:03
*** mxq has joined #crux11:10
*** lasso has joined #crux11:16
*** jjpk has joined #crux11:17
*** pitillo has quit IRC11:50
*** pitillo has joined #crux11:50
*** treach has joined #crux12:40
*** jaeger has quit IRC12:48
*** jaeger has joined #crux12:49
*** ChanServ sets mode: +o jaeger12:49
*** sepen_ has joined #crux13:09
*** sepen_ has quit IRC13:09
*** sepen has quit IRC13:09
*** sepen has joined #crux13:09
*** bjohan__ is now known as bjohan13:22
*** boe has joined #crux13:22
*** onestep has quit IRC13:33
rehabdolldoes anybody recognize this?: http://fredrik.obra.se/firefox.footprint13:33
rehabdollwierd change for just a version bump13:34
thrice`have you done a ports -u today ?13:38
*** fernan has joined #crux13:41
*** fernan has quit IRC13:42
jjpk"no activity?! quit quit quit quit!!!111" :p13:44
j^2hahah13:44
jjpkpush it a step forward and deduce crux is dead community-wise. :D13:46
jjpkHave to love this distribution's public profile.13:47
jjpk"yeah, we're dead, take a picture why don't you" :D :D13:48
RedShiftfact is13:49
RedShiftnobody cares13:49
jjpkSmall group of users, you can't expect much.13:51
RedShiftexactly13:52
RedShiftat least the crux community steps up to the users and says fuck you13:52
RedShiftif you don't like it take your business elsewhere13:52
RedShiftsomething archlinux is affraid to do13:53
RedShift(and I think in fear of losing popularity)13:53
jjpkIf the dead joke hit arch, it probably would be front page news on osnews13:53
jjpkIt's almost their pet.13:53
tilmanwe do?13:53
tilmanRedShift: we don't tell them to fuck themselves. we're delivering what crux always delivered13:54
tilmanand i think our users like it that way13:54
RedShiftyes that's true13:54
RedShiftbut one thing I've never understood is the need for a "games" group13:55
RedShiftI don't see what a game makes so different than any other application13:55
RedShiftyet, under community pressure... a games group was added13:55
RedShiftwell if we keep on sucking up that community pressure, we'll get another ubuntu13:56
teKbuying a voice regarding those decisions by commitment is a good thing, I think :)13:58
RedShiftteK: can you rephrase that?13:59
j^2RedShift: i think the games group was from back in the 70s on a unix mainframe where you could lock out the normal users14:01
RedShiftj^2: my point is, they added something so useless under community pressure14:01
j^2ahh14:02
RedShiftI wouldn't subcombe to community pressure, to hell with the community14:02
teKRedShift: one should be allowed to speek up (and be heard) if one showed commitment14:03
RedShiftagreed14:03
teKjust saying "a game repo" is not hard and the consequences  are only for the devs14:03
RedShiftteK: it's not a game repo14:04
RedShiftit's a game group14:04
RedShiftas in, a seperate group in /etc/groups14:04
tilmanteK: group as in /etc/group14:04
teKwoops ;P14:05
teKok. Useless as well in most cases14:05
*** mrks has joined #crux14:06
*** mrks_ has quit IRC14:19
*** mxq has left #crux14:28
*** muerto has quit IRC14:35
*** muerto has joined #crux14:36
*** onestep has joined #crux14:39
*** hp_tux has joined #crux15:14
teKanyone using SAS discs (with RAID-1) here?15:15
RedShiftteK: what about them?15:15
jaeger"anyone using" <-- this was the question part :)15:16
RedShiftwell I've sold machines with SAS controllers and SAS disks15:17
RedShiftand a few machines in my rack have SAS15:17
*** Alta_ve28 has quit IRC15:17
teKI've got 2 discs. reading via ftp/http over GBit Network gives me 50MB/s average.. is this ok?15:17
RedShiftthat seems slow15:21
RedShiftwhat kind of disks are they?15:21
RedShift10k rpm disks usually get about 80 megabytes per second15:21
teK15k.15:21
RedShiftthat should get you 100 megabytes per second15:21
teKdammit.15:21
*** Alta_ve28 has joined #crux15:21
RedShiftteK: what does hdparm give you?15:21
teK Timing buffered disk reads:  374 MB in  3.02 seconds = 124.05 MB/sec15:22
RedShiftso the problem lies somewhere else15:22
RedShiftwhat network card do you have?15:22
teKBroadcom Corporation NetXtreme BCM5715 Gigabit Ethernet (rev a3)15:22
RedShiftbroadcom support can be tricky at times15:23
RedShiftdo you have another gbit network card?15:23
teKon both sides (same HDDs, NIC,..); interconnected via cross15:23
RedShiftwhat if you wget them to devnull?15:23
RedShiftwhat speed do you get then?15:23
teKhave to try this. please standby15:23
teKuh oh..15:25
RedShiftstanding by15:25
teK00:22:39 (112.21 MB/s) - `/dev/null' saved [2147483602/2147483602]15:25
RedShift112 is about the maximum you'll get out of gigabit15:26
RedShiftso there's nothing wrong with your setup15:26
teKthe other machine seemed to have problems with its batter backed write cache.. maybe I should *recheck* this issue15:26
RedShiftteK: are you planning on using NFS with those machines?15:26
teKno. god beware.15:26
teKI'd love to do a 'full' backup of our ERP system during lunch *g*15:27
RedShiftwhat? something wrong with NFS?15:27
teKI think rsync will be a better choice15:27
teKI don't like NFS that much to be honest15:27
*** roliveira has joined #crux15:27
RedShiftwell I was going to give you a tip on NFS and gigabit, you have to use a bigger rsize and wsize15:28
RedShiftotherwise performance will suck and NFS will be hugely unreliable on gigabit15:28
teKI will try to remember that. thank you15:28
teKour hardware supporter suggested trying something else then FTP as "we had problems with FTP that often"15:28
RedShiftuse rsize=32768,wsize=32768 as options for mounting15:29
teKhis first question was "transfering with Explorer or what?"15:29
teK-> -_-15:29
RedShift(NFS default packet size is 1024, which causes huge amounts of fragmentation because packets are getting sent too fast)15:29
teKI see :)15:30
RedShift100 mb networks don't suffer that, although performance might increase if you use those parameters on 100mb networks as well15:30
jaegerhttp://stromberg.dnsalias.org/~strombrg/nfs-test.html15:30
teKthe idea with /dev/null should have come to my mind, too... the supporter said "50MB is fairly good, afaik" ..15:30
teKhopefully he meant using one disc, not two.15:31
RedShift50 MB is pretty weak with 15k rpm disks...15:31
RedShiftbecause a single 15k rpm disk can push at least 100 mb/s...15:31
teKholy..15:31
teKthe first server has a working BBU but: [ wait 2 more minutes ]15:35
teKroot@titus:/oradb1 # du -sh dumps15:35
teK7.9G    dumps15:35
teKroot@titus:/oradb1 # time cp -r dumps /oradb215:39
teKcp -r dumps /oradb2  0.40s user 23.04s system 10% cpu 3:39.84 total15:39
teKtwo separate RAID-1 sets, of two discs each..15:40
*** tri has joined #crux15:43
*** tri has left #crux15:43
rehabdollwill glibc 2.7 make it into 2.4?15:51
*** pitillo has quit IRC15:53
teKthank you so far, RedShift. Do have any idea why there could be this really bad disc-disc copy-performance?16:02
RedShifthmm 36 megabytes/sec is pretty poor16:02
RedShiftcaching set to write-through instead of write-back?16:02
*** jdolan_ has quit IRC16:03
teK@titus? No. It's write-through and write-back on the other machine with the broken(?) BBU16:04
teKno the other way round..16:04
RedShiftwell set it to write-back16:04
RedShiftand try again :D16:04
RedShiftyou can do that in the SAS controller's BIOS16:04
teKI checked this toady and it is write-back on the 'good' machine (titus)16:05
teKthe other machine gives me an error about the bbu being not charged or broken.16:05
RedShiftoh16:05
RedShiftnice16:05
RedShiftwell has it been charged, the battery?16:05
teKyep.16:05
RedShiftand it still says that?16:06
teKyep.16:06
RedShiftis the connection between the battery and the controller OK?16:06
teKthat's not too important right now, I will have it replaced or fixed..16:06
teKaccording to the BIOS it's present / charging and so on. I think so, yes.16:07
teKMore important to me is the fact that I get this poor performance despite write-back being active on the other machine16:07
*** mike_k has quit IRC16:08
teKeven worse: ftp FROM that machine TO /dev/null on the 'broken' one gives me 22:21 < teK> 00:22:39 (112.21 MB/s) - `/dev/null' saved [2147483602/2147483602]16:08
teKI don't get it.16:08
RedShiftftp from the broken machine?16:10
RedShiftwell if there is a battery problem, reading is never affected16:10
RedShiftthe battery is there to preserve the write cache, so if there's a problem with slow write speeds, that's where you would look (the write cache)16:10
teKtitus is the one with working BBU. _but_ it only performed with 36MB/s from disc to disc.16:11
teKthat's my problem.16:11
teKIt has write-back activated16:11
RedShiftah16:12
RedShiftwell that's pretty slow 36 MB/s16:12
RedShiftwat does hdparm -t give you on the first array and the second array?16:12
teKboth at about 124MB/s16:15
RedShiftok, so it's really writing that's a problem16:16
RedShiftI'd still inspect the controller's BIOS caching settings16:16
RedShiftwrite caching settings ofcourse16:16
*** jaeger has quit IRC16:17
teKok. I owe you a glass of beer something :-)16:18
teK+or16:18
RedShiftteK: it's solved?16:18
teKno but I had not much clue..16:19
teKIt's up to our supporter to help me tomorrow. (he does get paid)16:20
teKit 11pm and I got a 24x7 contract plus his mobilenr. ... :>16:21
RedShifthehehe16:21
teKone last thing (YFYI): broken-bbu/controller machine to "good" machine(titus):  23:20:26 (69.62 MB/s) - `/dev/null' saved [8589934378/8589934378]16:24
teKthe controller must have some problems ... broken battery should not affect reading..16:25
treachrehabdoll: yeah, I get the same footprint error, got it for the previous version too.16:26
*** boe has quit IRC16:28
teK< get some sleep. n8.. and thx again RedShift16:28
RedShiftno problem16:29
*** jaeger has joined #crux16:47
*** ChanServ sets mode: +o jaeger16:47
*** onestep has quit IRC16:52
*** onestep has joined #crux16:52
*** hp_tux has quit IRC16:59
*** mwansa has quit IRC17:06
*** lasso has quit IRC17:11
*** mwansa has joined #crux17:12
*** onestep has quit IRC17:15
*** muerto has quit IRC17:18
*** RedShift has quit IRC17:20
*** mwansa has quit IRC17:27
*** mwansa has joined #crux17:44
mwansarehabdoll, did you find out if glibc 2.7 is gonna get pushed into 2.7 ?17:44
mwansa2.4*17:45
*** jjpk has quit IRC17:56
*** jmvr has joined #crux18:07
*** jmvr has quit IRC18:08
*** _mavrick61 has quit IRC19:12
*** Dudde has quit IRC19:12
*** _mavrick61 has joined #crux19:13
*** Dudde has joined #crux19:13
*** predatorfreak has joined #crux19:15
*** roliveira has quit IRC19:19
*** predatorfreak has quit IRC19:39
*** djfp has joined #crux20:20
djfpAnyone use crux ppc?20:20
treachyeah, but they aren't here.20:22
treachhmm, "alasctl restore". No wonder that didn't work. :>20:24
*** schniggie has quit IRC21:33
*** thrice` has quit IRC22:54
*** djfp has quit IRC22:56
*** mwansa has quit IRC23:01
*** djfp has joined #crux23:02
*** morlenxus has quit IRC23:09
*** morlenxus has joined #crux23:09
*** djfp has quit IRC23:10
*** thrice` has joined #crux23:11
*** jaeger has quit IRC23:27
*** mwansa has joined #crux23:35

Generated by irclog2html.py 2.11.0 by Marius Gedminas - find it at mg.pov.lt!