1125187205 Q * Doener Ping timeout: 480 seconds 1125187206 J * Doener ~doener@p548779EB.dip.t-dialin.net 1125191402 N * Bertl_oO Bertl 1125191410 M * Bertl evening folks! 1125191569 M * Bertl phew, didn't think I would make it today ... my isdn router kicked the bucket :/ 1125193118 M * Bertl okay, folks, I'm off to bed now ... back later today ... 1125193163 N * Bertl Bertl_zZ 1125194256 Q * keyser_soze Quit: Abandonando 1125195735 Q * lilo_ Ping timeout: 480 seconds 1125195860 J * lilo tor@lilo.usercloak.oftc.net 1125201198 J * dddd44 ~dhb55@lostpackets.de 1125201550 M * dddd44 excuse me .. why the proxy server divided into 2 type: http and socks?? why not use http proxy for every application?? can't i use http for irc?? why i must use sock in irc??anyone?? 1125206988 J * kas_3 ~dhb55@tor-irc.dnsbl.oftc.net 1125207415 Q * dddd44 Ping timeout: 480 seconds 1125209312 Q * jkl_ Quit: BitchX-1.1-final -- just do it. 1125209749 Q * kas_3 Quit: Leaving 1125219365 Q * Greek0 Read error: Connection reset by peer 1125219504 J * Greek0 ~greek0@81.189.246.175 1125221722 Q * Greek0 Remote host closed the connection 1125221799 J * renihs ~renihs___@193.170.52.70 1125223103 J * Greek0 ~greek0@81.189.246.175 1125225423 N * _ag_ a-g 1125225445 N * a-g _ag_ 1125225538 N * _ag_ ag 1125225552 N * ag _ag_ 1125225830 J * PalleLap ~bo@50C56B02.flatrate.dk 1125225848 N * _ag_ ag- 1125226104 N * ag- ag-2 1125226131 J * yarihm ~yarihm@217-162-114-40.dclient.hispeed.ch 1125232000 Q * Nicoli Quit: ( www.nnscript.de :: NoNameScript 3.81 :: www.XLhost.de ) 1125234971 N * ag-2 ag- 1125235040 N * ag- _ag_ 1125235123 N * _ag_ ag- 1125235149 N * ag- ag-2 1125235184 N * ag-2 ag- 1125235571 Q * ag- Quit: Reconf... 1125235716 J * ag- ag@caladan.roxor.cx 1125236340 Q * ag- Quit: leaving 1125237150 J * ag- ag@caladan.roxor.cx 1125237188 Q * ag- Quit: 1125237201 J * ag- ag@caladan.roxor.cx 1125237230 J * ntrs ~ntrs@d60-65-131-192.col.wideopenwest.com 1125237268 M * ntrs does anyone know how to change the output of uname -a inside a virtual server? What is the config file I need to change? 1125237292 Q * ag- Quit: 1125237417 M * Hollow ntrs: /etc/vservers/name/uts 1125237441 M * ntrs yes, but which file in uts? 1125237461 M * ntrs found it. 1125237462 M * ntrs Thanks. 1125237467 Q * ntrs Quit: 1125237489 J * ag- ag@caladan.roxor.cx 1125237845 Q * nokoya Quit: changing servers 1125237860 Q * ag- Quit: leaving 1125237883 J * nokoya young@hi-230-82.tm.net.org.my 1125238035 Q * SiD3WiNDR Ping timeout: 480 seconds 1125238078 J * ag- ag@caladan.roxor.cx 1125240046 A * Hollow joggles Bertl_zZ to awaken him 1125240059 J * Nicoli ask@208.53.159.170 1125240922 N * Bertl_zZ Bertl 1125240962 M * Bertl moring folks! Hollow! 1125240966 M * Hollow :) 1125241001 M * Hollow had a question wrt iattr, but seems like i found it myself ;) 1125241030 M * Bertl (the answer, I presume :) 1125241049 M * Hollow yep ;) 1125241208 M * Hollow Bertl: in which situation would i need the dump_history syscall? 1125241234 M * Bertl when you want to know what happened until now (modulo the size of the history buffer) 1125241244 M * Bertl s/until/up to/ 1125241266 M * Hollow couldn't i just use dmesg? 1125241287 M * Bertl the history buffer is not dumped unconditional 1125241302 M * Bertl it is dumped when there is a kernel oops/panic 1125241310 M * Bertl so you could alternatively do that :) 1125241362 M * Hollow k 1125241523 M * Hollow Bertl: and i have no idea what get_rlimit_mask is about 1125241611 M * Bertl the kernel _knows_ what limits are supported 1125241627 M * Bertl to make that extensible, it reports this info back when asked 1125241653 M * Bertl so you basically 'should' first check what limits _are_ there, then query and/or set them 1125241657 M * Hollow ah, this comes in handy :) 1125242003 M * Nicoli mornin 1125242018 M * Bertl hey Nicoli! everything fine? 1125242053 M * Nicoli yea 1125242147 M * Bertl Nicoli: so you still want to 'donate' 100 USD for some help with linux-vserver? 1125242317 M * Nicoli yea 1125242401 J * SiD3WiNDR luser@bastard-operator.from-hell.be 1125242409 M * Bertl evening SiD3WiNDR! 1125242458 M * Bertl Nicoli: okay, in 2 hours is fine for you? 1125242475 M * Nicoli ok 1125242704 Q * Loki|muh_ Ping timeout: 480 seconds 1125243018 J * Loki|muh loki@satanix.de 1125243316 M * Hollow Bertl: http://home.xnull.de/misc/tools.html many things already work like a charm ;) 1125244031 M * Bertl okay, back after dinner ... 1125244044 N * Bertl Bertl_oO 1125244309 J * keyser_soze ~cimarron@host245.201-252-8.telecom.net.ar 1125245128 Q * Greek0 Quit: Lost terminal 1125245189 J * Greek0 ~greek0@81.189.246.175 1125245830 J * dddd44 ~dhb55@tor-irc.dnsbl.oftc.net 1125245862 M * dddd44 excuse me what is my ip address? can you ping??anyone? 1125245985 M * DaCa dddd44: the ip of your tor exit node of course :) 1125246810 Q * tchan Remote host closed the connection 1125246868 J * tchan ~tchan@c-24-13-81-164.hsd1.il.comcast.net 1125247002 Q * micah Quit: leaving 1125247017 J * micah micah@micha.hampshire.edu 1125247168 Q * micah Quit: 1125247183 J * micah micah@micha.hampshire.edu 1125247269 J * stefani ~stefani@c-24-19-46-211.hsd1.wa.comcast.net 1125247444 N * Bertl_oO Bertl 1125247452 M * Bertl back now ... 1125247672 M * Greek0 Bertl: do you have time for some discussion about the rlimit stuff? 1125247910 M * Bertl guess so ... 1125247994 M * Nicoli heh 1125248053 M * Greek0 how do you actually imagine a softlimit implementation? I mean how would contexts which exceed the softlimit be penalized? 1125248081 M * Bertl for example by removing tokens from the TBS? 1125248102 M * Greek0 whenever it exceeds _any_ softlimit? 1125248125 M * Bertl depends on the softlimit 1125248173 M * Greek0 mm 1125248209 M * Bertl for example, let's take the 'fork' limit 1125248211 M * Greek0 well, I could imagine that well for eg. the NPROC limit 1125248219 M * Bertl yeah :) 1125248240 M * Bertl we could effectively 'delay' a fork for some time ... 1125248273 M * Bertl could even make it exponential depending on how far it is beyond the soft limit 1125248296 M * Greek0 mm 1125248335 M * Greek0 well, what about the memory limits (RSS, AS, MEMLOCK, ..)? 1125248363 M * Bertl rss limits could be used for classifying page eviction 8swaping) 1125248390 M * Bertl i.e. if the system (kernel) decides to swap out pages, they could get a priorization regarding to those limits 1125248411 M * Greek0 ah, nice 1125248663 M * Greek0 hmm well. so what do you want softlimits to be conceptually? some sort of penalty if the system is low on resources, or a constant penalty (even if resources would be available)? 1125248683 M * Greek0 and how does that interact with the minlimits then? 1125248722 M * Bertl I guess we will have to find out, basically by testing ... 1125248741 M * Bertl I'm not convinced that the min/soft limits will make sense in every rlimit 1125248750 M * Greek0 yep 1125248765 M * Bertl and I'm completely sure that their purpose will be different from one limit to the other 1125248790 M * Bertl especially as there _are_ resources which can be granted but not withdrawn 1125248823 M * Greek0 so the first step would be to actually track them in the kernel (i.e. update _vx_limit to include them and make the limit functions aware of these changes) 1125248833 M * Bertl for example, the number of filehandles, you can limit that, but a) what to do if over soft limit? and b) what to do when the hard limit should be lowered? 1125248872 M * Bertl Greek0: the first required hanges are neglectible, they are just _allowing_ the soft/min to be set, nothing else 1125248877 M * Bertl *changes 1125248898 M * Bertl (everything for that is already in place, it's only a few lines of changes) 1125248924 M * Bertl the next step, could be to _warn_ on soft limits .. like quota warnings 1125248938 M * Greek0 mm 1125248940 M * Bertl first only for the host/admin, later for the guests (if selected) 1125248989 M * Hollow .oO( context state changes ) 1125249130 Q * cryo Ping timeout: 480 seconds 1125249138 M * Bertl Nicoli: okay, let's get started with your system, please msg me the required access data 1125249145 M * Nicoli oki doki 1125250462 Q * dddd44 Quit: Leaving 1125251932 M * Bertl Greek0: for the min values some kind of non-overbooking check would be useful (at set time) 1125251963 M * Bertl for example, folks tend to overbook the semaphores and memory stuff ... the min-limit could warn there on setup ... 1125252037 M * Greek0 aeh, overbook? 1125252048 M * Bertl Hollow: maybe rlimit set operation should also support some kind of 'not available, reduced to .. please reread values' return code? 1125252080 M * Bertl Greek0: that's what hotels and flight companies do with places :) 1125252096 M * Hollow Bertl: kernel, lib, tools? 1125252109 M * Bertl kernel return code, lib/tools support 1125252134 M * Bertl we could also do it the other way round, have some syscall command to query available resources 1125252163 M * Hollow well, the rlimits are implemented as a list_t with only valid values, so the lib doesn't allow usage of non-valid rlimit ids 1125252247 M * Hollow http://dev.croup.de/proj/libvserver/file/trunk/util/rlimit-list.c 1125252269 M * Hollow btw.. the VLIMIT_* defines should be for userspace to, right? 1125252430 M * Greek0 Bertl: somehow I fail to see what overbooking has to do with min limits. 1125252835 J * cryo ~say@212.86.243.154 1125253137 M * Bertl wb cryo! 1125253166 M * Bertl Hollow: yep, they are for userspace too ... 1125253185 M * Hollow Bertl: they're included in __kernel__ ifdefs... 1125253291 M * Bertl not in my kernel sources ... 1125253310 M * Hollow hmm 1125253331 M * Hollow maybe i should update those in the lib.. ;) 1125253379 M * Bertl yep, update to vs2.0 would be great! :) 1125253393 M * Hollow :P 1125253436 M * Bertl btw, Greek provided a mips syscall implementation, we will integrate that into shiny soon 1125253443 M * Bertl *Greek0 1125253444 M * Hollow Bertl: btw, it seems without chaining there is no possibility currently to set both xid and nid of a process 1125253452 M * Hollow hurray :) 1125253483 M * Bertl xid and nid in one go, you mean 1125253521 M * Hollow it doesn't work in seperate goes as well, at least i wasn't able to do it 1125253539 M * Bertl hmm, why? 1125253551 M * Hollow EPERM 1125253561 M * Hollow with chaining 1125253562 M * Hollow and without 1125253570 M * Hollow where should i specify the xid? ;) 1125253800 M * Hollow hm 1125253817 M * Hollow it works the other way round.. create nx first, then chain with vx create 1125254098 M * Bertl yup 1125254164 M * Bertl okay, I'm off for about 20 minutes, have to configure the replacement router ... 1125254174 N * Bertl Bertl_oO 1125255862 M * Nicoli ... 1125256101 P * stefani parting (is such sweet sorrow) 1125257187 J * mess-mate ~mess-mate@lns-vlq-7-lil-82-254-205-79.adsl.proxad.net 1125258094 J * mess-mat1 ~mess-mate@lns-vlq-7-lil-82-254-205-32.adsl.proxad.net 1125258116 J * monski ~monrad@213083190134.sonofon.dk 1125258128 Q * mess-mate Read error: Operation timed out 1125258254 Q * mess-mat1 Quit: 1125258418 N * Bertl_oO Bertl 1125258430 M * Bertl okay, back now ... did take a little longer than expected 1125258435 Q * yarihm Quit: Leaving 1125258492 Q * keyser_soze Remote host closed the connection 1125258545 Q * monrad Ping timeout: 480 seconds 1125259709 Q * lilo Read error: Connection reset by peer 1125259822 M * Aiken pre5 will still oops with http://pastebin.com/348633 1125259875 M * Bertl ah, morning Aiken! 1125259884 M * Bertl seems we missed yesterday ... 1125259887 M * Aiken hello 1125260045 M * Bertl sec, let me recreate this ... 1125260072 M * Aiken in a way it was nice spending the day outside instead of infront of this thing :) 1125260109 M * Bertl sounds like you need wireless and a laptop :) 1125260139 M * Bertl but you should know that the testing and debugging you do is very appreciated! 1125260165 M * Aiken http://pastebin.com/348638 1125260187 M * Aiken I don't mind. I am hoping it is at least putting something back 1125260212 M * Aiken I have 5 vservers runnings spread across 3 machines 1125260273 M * Bertl actually it is one of the most valuable inputs .. because a bug I know of ... can be fixed :) 1125260367 M * Bertl Aiken: hmm, my current version works quite fine with your commands 1125260379 J * lilo ~lilo@lilo.usercloak.oftc.net 1125260404 M * Aiken Linux pebbles.bedrock 2.6.13-rc7-vs2.1.0-pre5 #1 Fri Aug 26 15:44:21 EST 2005 alpha Unknown Alcor GNU/Linux 1125260430 M * Aiken that is not what I wanted you to say 1125260481 M * Bertl hmm ... did you build it froma source aptched with the all-in-one patch or did you use the delta? 1125260530 M * Aiken I ended up building with the pre5 patch 1125260538 M * Bertl hmm .. k 1125260626 M * Aiken should namei.c be the same either way? 1125260724 M * Aiken after the other day the tree was in bad enough shape it was easiest to start again so just went pre5 1125260840 M * Bertl let me verify the patch ... (i.e. I make another one and compare them with the uploaded) 1125260973 M * Bertl okay, except for an empty line (which is now removed) my current source is dientical with the pre5 (doing another check for the base kernel) 1125261149 M * Aiken this is the oops I did this morning http://pastebin.com/348638 1125261176 M * Bertl looks very similar ... 1125261538 M * Bertl okay, we are going to do two things ... a) I've asked my friend to power up the alpha, so I can test it on alpha .. and b) we should try to 'redo' it step by step in sync (first on x86/alpha, if that doesn't show any differences, then on alpha/alpha) okay? 1125261576 M * Aiken ok 1125261653 M * Bertl okay, what filesystem do you do the http://pastebin.com/348633 test on? 1125261661 M * Aiken ext3 1125261665 M * Aiken local filesystem 1125261692 M * Bertl okay, normal journal? 1125261716 M * Bertl can you mount it anywhere? 1125261737 M * Aiken mkfs.ext3 /dev/sda2 is the cmd used 1125261741 M * Aiken it is / 1125261800 M * Bertl hmm, okay, I have mkfs.ext3 /dev/discs/disc1/part1 and I'll mount it on /mnt 1125261858 M * Bertl okay, so I'll do the commands with pwd= /mnt and you with /, right? 1125261870 M * Aiken yes 1125261892 M * Bertl okay, you're ready, can we try again with some check inbetween? 1125261916 M * Aiken I am ready 1125261920 M * Bertl mkdir a 1125261937 M * Bertl echo blah > a/test 1125261952 M * Aiken done 1125261958 M * Bertl ln a/test 1125261971 M * Aiken done 1125261976 M * Bertl ls -la a test* 1125261985 M * Bertl -rw-r--r-- 2 root root 5 Aug 28 20:42 test 1125261985 M * Bertl a: 1125261985 M * Bertl total 3 1125261985 M * Bertl drwxr-xr-x 2 root root 1024 Aug 28 20:42 . 1125261985 M * Bertl drwxr-xr-x 4 root root 1024 Aug 28 20:42 .. 1125261987 M * Bertl -rw-r--r-- 2 root root 5 Aug 28 20:42 test 1125262027 M * Aiken (root@pebbles) ls -la a test* 1125262027 M * Aiken -rw-r--r-- 2 root root 5 2005-08-29 06:45 test 1125262027 M * Aiken a: 1125262027 M * Aiken total 12 1125262027 M * Aiken drwxr-xr-x 2 root root 4096 2005-08-29 06:45 . 1125262028 M * Aiken drwxr-xr-x 22 root root 4096 2005-08-29 06:46 .. 1125262029 M * Aiken -rw-r--r-- 2 root root 5 2005-08-29 06:45 test 1125262038 M * Bertl k 1125262058 M * Bertl echo qwerty > test1 1125262078 M * Bertl setattr --iunlink test 1125262101 M * Aiken setattr --iunlink test 1125262110 M * Aiken try that in the correct window 1125262125 M * Bertl showattr a test* 1125262125 M * Bertl ---bui- a 1125262125 M * Bertl ----UI- a/test 1125262125 M * Bertl ----UI- test 1125262166 M * Aiken (root@pebbles) showattr a test* 1125262166 M * Aiken ---bui- a 1125262166 M * Aiken ----UI- a/test 1125262166 M * Aiken ----UI- test 1125262166 M * Aiken lstat(): No such file or directory 1125262176 M * Bertl hmm ... 1125262177 M * Aiken that last bit is wierd 1125262187 M * Bertl ah, no, I have it too ... 1125262199 M * Bertl just didn't copy it (no idea why :) 1125262237 M * Aiken here is an strace of it http://pastebin.com/348662 1125262244 M * Aiken lstat is complaining about test1 1125262271 M * Bertl yeah, funny 1125262334 M * Bertl I suspect a tool bug somewhere ... 1125262345 M * Bertl but maybe it's a kernel issue ... sec 1125262382 M * Aiken but showattr test* a works 1125262403 M * Bertl yeah, it's weird 1125262423 M * Aiken (root@pebbles) showattr test* a 1125262423 M * Aiken ----UI- test 1125262423 M * Aiken ----ui- test1 1125262423 M * Aiken ---bui- a 1125262423 M * Aiken ----UI- a/test 1125262435 M * Bertl lsattr a test* 1125262435 M * Bertl ----i-------- a/test 1125262435 M * Bertl ----i-------- test 1125262435 M * Bertl ------------- test1 1125262445 M * Bertl I guess that works for you too, no? 1125262456 M * Bertl Greek0, Hollow: still around? 1125262461 M * Greek0 . 1125262464 M * Hollow yup 1125262485 M * Aiken (root@pebbles) lsattr a test* 1125262485 M * Aiken ----i-------- a/test 1125262485 M * Aiken ----i-------- test 1125262485 M * Aiken ------------- test1 1125262488 M * Bertl okay, we had some kind of tool issue with gcc/syscall no? 1125262514 M * Hollow did we? 1125262521 M * Bertl with the older shiny used in the tools ... where some syscall was overwritten? 1125262539 M * Bertl (not syscall, but the version syscall did overwrite the ebx register= 1125262551 M * Hollow yeah, i remember 1125262563 M * Bertl Aiken: what tools do you use? and what patches if any (for the tools)? 1125262609 M * Aiken vserver tools or the main toolchain? 1125262616 M * Aiken vserver tools are 208 1125262625 M * Bertl util-vserver 0.30.208 I assume, but any patches? 1125262632 M * Aiken no patches 1125262678 M * Bertl okay, doing two cross check now .. a) with latest patches, and b) with the vs2.0 kernel 1125262839 M * Aiken do you want the final copy done with will cause the oops or just wait? 1125262846 M * Bertl not yet 1125262927 M * Bertl okay, it fails with 0.30.204 or 208(not latest) on vs2.0 too ... funny ... 1125262953 M * Bertl now testing latest tools+patches (compiling) 1125262963 M * Aiken built a cross tool chain a couple of days ago so I can now compile an alpha kernel in 12 min instead of > 70 1125262973 M * Aiken from scratch 1125262984 M * Bertl excellent! 1125263201 M * Hollow 208 is not latest? 1125263235 M * Bertl 208 + fix02 is :) 1125263279 M * Bertl Aiken: so http://vserver.13thfloor.at/Experimental/patch-2.6.13-rc7-vs2.1.0-pre5.diff.bz2 is what you used (or the diff) right? 1125263306 M * Aiken yes 1125263892 M * Bertl okay, let's continue with the tests ... 1125263920 M * Bertl cp test1 test 1125263927 M * Bertl (works fine here) 1125263987 M * Aiken oops 1125264074 M * Aiken fs/namei.c:2568 1125264132 M * Aiken (root@fred) md5sum fs/namei.c 1125264133 M * Aiken df15672f5a44b5c3e64ecccf968abca1 fs/namei.c 1125264165 M * Bertl md5sum fs/namei.c 1125264166 M * Bertl 686d974b9f938808c9894076105121ad fs/namei.c 1125264180 M * Bertl interesting ... 1125264192 M * Bertl let me upload it ... 1125264263 M * Bertl http://vserver.13thfloor.at/Stuff/fs_namei.c 1125264303 M * Aiken VERSION = 2 1125264304 M * Aiken PATCHLEVEL = 6 1125264304 M * Aiken SUBLEVEL = 13 1125264304 M * Aiken EXTRAVERSION = -rc7-vs2.1.0-pre5 1125264399 M * Bertl rc7-vs2.1.0-pre5 we agree on that :) 1125264427 M * Aiken the difference is superficial http://pastebin.com/348685 1125264458 M * Aiken a blank line and whitespace 1125264461 M * Bertl okay ... 1125264469 M * Bertl this gets removed by my patch cleaner ... 1125264692 M * Aiken what gcc are you using? 1125264726 M * Bertl 3.3.5/6 for the kernel 1125264748 M * Aiken a std 3.3.6 1125264796 M * Bertl is that a question? :) 1125264834 M * Aiken statement, I am using a std 3.3.6 1125264840 M * Bertl k 1125264941 M * Bertl hmm, what is at line 2568 in your fs/namei.c? 1125264966 M * Bertl (and di you do the echo 255 >/proc/sys/vserver/debug_misc before?) 1125264995 M * Aiken 2568 -> 1125264995 M * Aiken vxdprintk(VXD_CBIT(misc, 3), 1125265012 M * Aiken that echo is in rc.local so it is always set 1125265042 M * Bertl okay, let me try with that on :) 1125265070 M * Aiken http://pastebin.com/348694 1125265129 M * Hollow Bertl: anything special with the ebx thing? cause i might be off to bed now 1125265171 M * Bertl Hollow: no, with fix2 (or shiny7) it's fine 1125265178 M * Bertl (at least we think so :) 1125265179 M * Hollow k 1125265226 M * daniel_hozac Bertl: src/fstool.c, processFile chdir's to the argument if it's a directory, without chdir'ing back before returning. 1125265270 M * Bertl ah, that sounds interesting ... 1125265291 M * Bertl so it's a tool bug ... actually 1125265494 M * Bertl hmm, it doesn't fail here (with debugging, but the output looks fishy) 1125265503 M * Bertl [ 418.842833] vxD: fput(new_file=8f050f5c[#1]) 1125265503 M * Bertl [ 418.845335] vxD: fput(old_file=8f04ff5c[#1]) 1125265503 M * Bertl [ 418.847631] vxD: dentry@end: 8f747000[#1953719668] 1125265521 M * Bertl so I suspect the dentry was invalidated and dropped in the meantime 1125265580 M * Bertl Aiken: please try to disable debugging for a test? or remove the last vxdprintk() with the dentry@end completely 1125265782 M * Aiken no crash 1125265785 M * Aiken it worked 1125265813 M * Aiken echo 0 >/proc/sys/vserver/debug_misc to disable debug 1125265831 M * Bertl okay, I assume we just release it accidentially ourself, but I will review the code once again ... 1125265918 M * Bertl if you plan to do extensive testing, please remove that vxdprint and watch for increasing dentries and/or filesystem which cannot be unmounted or similar 1125266243 M * Aiken with debug off I just started a cow with no oops 1125266351 M * Bertl excellent ... 1125266360 M * Aiken 2nd cow started ok 1125266509 M * Aiken cows 3 and 4 started ok 1125266535 M * Bertl so you are running 4 completely unified guests, right? 1125266578 M * Aiken yes 1125266611 M * Bertl cool! 1125266799 M * Aiken I am now tempted to go back to using aoe to get this lot off / 1125266824 M * Aiken have 2 of the guests working, one was entered locally, the other is ssh in 1125266840 M * Aiken system load abt 2.7 1125266881 M * Bertl okay, I'd suugest you compile a kernel without the dentry vxdprint() 1125266888 M * Bertl *suggest 1125266901 M * Aiken already commented out 1125266902 M * Bertl and go crazy with the testing over aoe ... 1125266933 M * Aiken I am starting to prefer aoe over iscsi 1125266987 M * Bertl yeah, sounds interesting .. I guess I have to have a look at it soon ... 1125266988 M * Aiken with iscsi I can easily make the server (P3-600) go 100% cpu and with aoe the server maxs out at 20%cpu 1125267022 M * Bertl how is the performance compared to nfs (except for the sharing which isn't really possible) 1125267042 M * Bertl well, not without a distributed filesystem at least 1125267063 M * Aiken I have found both aoe and iscsi faster than nfs 1125267079 M * Aiken one think you get with the network block device is caching which you don't have with nfs 1125267100 M * Aiken dealing with block device instead of a filesystem 1125267129 M * albeiro is there any iscsi or aoe _server_ implementation for linux ? 1125267137 M * albeiro so far i have seen only clients 1125267219 M * Aiken service side I have been using http://sourceforge.net/project/showfiles.php?group_id=108475 with iscsi 1125267231 M * Aiken and for aoe a package called vblade 1125267262 M * Bertl http://dcs.nac.uci.edu/~strombrg/AoE-notes.html 1125267267 M * Aiken for the iscsi target I stayed 0.3.8 as that is the last version that supports a 2.4 kernel 1125267299 M * Aiken iscsi on the client I use http://linux-iscsi.sourceforge.net/ 1125267335 M * Aiken I am finding aoe marginally faster than iscsi and a lot less load on the server 1125267351 M * Aiken though the iscsi server load could be the implementation I am using 1125267410 M * albeiro i wonder what could be better 1125267419 M * albeiro aoe or iscsi 1125267442 M * albeiro ok, just have read back ;) 1125267492 M * Greek0 ok, I'm off to bed, cu 1125267499 M * Bertl night Greek0! 1125267521 M * daniel_hozac Bertl: http://daniel.hozac.com/stuff/util-vserver-0.30.208-fstool-dir.patch seems to fix that problem, look ok? 1125267523 A * albeiro about to install debian @ ppc . wish me luck ;) 1125267806 M * Bertl daniel_hozac: looks good, I assume you tested it? 1125267806 Q * PalleLap Read error: Connection reset by peer 1125267841 M * Bertl albeiro: should be a piece of cake :) 1125267849 M * renihs me is off too for awhile, night Bertl! 1125267850 Q * renihs Remote host closed the connection 1125267859 M * daniel_hozac Bertl: yep, solves the problem here. 1125267887 M * Bertl daniel_hozac: okay, great, thanks! could you submit it to savannah too if not already done so? 1125267927 M * daniel_hozac just wanted to make sure i wasn't on crack before submitting it ;) 1125267950 M * Bertl btw, you have a bunch of patches for the fs tools/coreutils, IIRC, no? 1125267965 M * albeiro Bertl: that's how it looks for now. been using it for a long time, but a while ago. 1125267974 M * daniel_hozac well, a few patches. 1125267987 M * daniel_hozac findutils and coreutils. 1125267994 M * Bertl daniel_hozac: what about if we do some 'official' aptch collection for that? 1125268009 M * daniel_hozac as in? 1125268026 M * Bertl no idea, jsut thinking out loud ... 1125268036 M * Bertl maybe a page on 13thfloor (as addons or so?) 1125268060 M * Bertl or a wiki page with a short explanation why and what? 1125268071 M * daniel_hozac IIRC it's already on the wiki 1125268097 M * Bertl hmm .. *looking* 1125268108 M * daniel_hozac http://linux-vserver.org/Tools+and+patches 1125268111 M * daniel_hozac last section. 1125268198 M * Bertl ah, okay, great, you already have them together on one page ... 1125268209 M * Bertl perfect, please keep up the good work! 1125268290 M * daniel_hozac thanks. 1125268303 M * Bertl btw, I thought about a really minimalistic (maybe dietlibc?) based host distro with a lot of patched tools, what do you folks think about that? 1125268346 M * daniel_hozac will the minimalism work with all the build methods? 1125268360 M * daniel_hozac yum at least has quite a bit of dependencies. 1125268390 M * Bertl maybe there are different requirements to satisfy ... 1125268401 M * Bertl I can imagine a host system based solely on templates 1125268467 M * Bertl iyou could then 'build' your templates on a different machine 1125268481 M * Aiken sounds like one of the vmware products, enough to boot the host and run many guest 1125268528 M * Bertl don#t know it, but yes, that's the idea ... with a lot of useful but vserver specific tools 1125268553 M * Bertl maybe even some monitoring (rrd and such) 1125268563 M * Bertl maybe with a webfrontend in the future? 1125268573 M * Aiken I think it is a good idea 1125268580 M * daniel_hozac yeah, sounds cool. 1125268663 M * Bertl because I have the feeling that the over all experience/quality is not that high, when we try to support every distro (including latest test stuff) perfectly ... 1125268687 M * Bertl such a special distro would allow to steadily improve it ... 1125268726 M * Bertl and I don#t think we would suffer from exploits/security stuff that much, because we will not provide services other than ssh and maybe the web server on the host 1125268748 M * Bertl (and I'm not sure about the web server :) 1125268798 M * Bertl well, for now it's just an idea ... and suggestions for existing distros which would be a good base for that (including LFS :) are welcome 1125268852 A * Aiken grins at the idea of LFS 1125268882 M * Aiken which would be the easiest you to build a host exactly as you want? 1125268992 M * Aiken you mentioned dietlibc so wouldn't you have to build something hand anyway? 1125269029 M * Bertl yes, probably ... unless there is a distro based on dietlibc :) 1125269135 M * daniel_hozac bug #14313 submitted. 1125269146 M * daniel_hozac it appears savannah doesn't escape < and > though... 1125269213 M * Aiken Bertl http://www.dietlinux.de/ 1125269400 M * Bertl looks good ... 1125269480 M * Bertl seems I already looked at that two years ago :) 1125269494 M * Bertl 20467712 Jun 2 2003 dietlinux-snapshot.iso 1125269529 M * Bertl so we can consider it mature too ... *G* 1125269886 N * zobel devscripts 1125269905 N * devscripts zobel 1125269906 M * Bertl devscripts? 1125270208 J * Aiken_ ~james@tooax7-081.dialup.optusnet.com.au 1125270260 Q * stupidawy Read error: Operation timed out 1125270292 Q * Aiken Remote host closed the connection 1125270293 Q * tchan Read error: Operation timed out 1125270314 N * Aiken_ Aiken 1125270539 Q * monski jupiter.oftc.net oxygen.oftc.net 1125270539 Q * cryo jupiter.oftc.net oxygen.oftc.net 1125270539 Q * Greek0 jupiter.oftc.net oxygen.oftc.net 1125270539 Q * menomc jupiter.oftc.net oxygen.oftc.net 1125270539 Q * virtuoso jupiter.oftc.net oxygen.oftc.net 1125270539 Q * eyck jupiter.oftc.net oxygen.oftc.net 1125270539 Q * albeiro jupiter.oftc.net oxygen.oftc.net 1125270654 J * Greek0 ~greek0@81.189.246.175 1125270691 J * stupidawy foo@198.77.239.131 1125270731 J * monski ~monrad@213083190134.sonofon.dk 1125270760 J * mnemoc ~amery@200.75.27.38 1125270886 J * virtuoso ~s0t0na@80.253.205.251 1125270886 J * tchan ~tchan@c-24-13-81-164.hsd1.il.comcast.net 1125270886 J * cryo ~say@212.86.243.154 1125270886 J * albeiro ~albeiro@graffias.estrefa.pl 1125270960 J * eyck eyck@81.219.64.71 1125271181 M * Aiken 2 cows running, both source based, both going to spend the next hour or more updating 1125271736 Q * albeiro helium.oftc.net oxygen.oftc.net 1125271736 Q * cryo helium.oftc.net oxygen.oftc.net 1125271736 Q * tchan helium.oftc.net oxygen.oftc.net 1125271736 Q * virtuoso helium.oftc.net oxygen.oftc.net 1125271739 J * Aiken_ ~james@tooax6-122.dialup.optusnet.com.au 1125271858 J * tchan ~tchan@c-24-13-81-164.hsd1.il.comcast.net 1125271890 J * albeiro ~albeiro@graffias.estrefa.pl 1125272067 Q * Aiken Ping timeout: 480 seconds 1125272144 J * virtuoso ~s0t0na@80.253.205.251 1125272145 J * cryo ~say@212.86.243.154 1125272286 J * lilo_ ~lilo@lilo.usercloak.oftc.net 1125272400 Q * lilo Ping timeout: 480 seconds 1125272643 Q * cryo jupiter.oftc.net oxygen.oftc.net 1125272844 J * cryo ~say@212.86.243.154 1125272937 J * renihs ~renihs___@193.170.52.70