T2 IRC Log: 2007-11-10
This is the log as captured by an IRC bot in the channel. The statements are those of the individual people and might not neccessarily reflect the policy and legal rules as set forth by the T2 SDE Project.
--- Log opened Sat Nov 10 00:00:15 2007
05:07 -!- yokoy [n=yokoy@e178077202.adsl.alicedsl.de] has joined #t2
05:21 -!- yokoy_ [n=yokoy@e178075071.adsl.alicedsl.de] has quit [Read error: 110 (Connection timed out)]
08:13 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has quit ["There is intelligent life out there..."]
10:19 < rxr> re
10:19 < rxr> Ragnarin: which libtool build failure are you looking at ?
10:19 < Ragnarin> see the last few commits
10:19 < rxr> Ragnarin: the libtool hackery often is a new jungle, I think there is nothing like "not this dir"
10:19 < Ragnarin> wasn't actually libtools fault
10:19 < rxr> oh
10:19 < rxr> cia where are you ????
10:19 < Ragnarin> the package had a -L/usr/lib hardcoded
10:20 < rxr> finally:
10:20 < rxr> Error logs from default-7.0-stable-generic-sparc64-generic:
10:20 < rxr> 231 builds total, 231 completed fine, 0 with errors, 0 to go.
10:33 < rxr> hm - even on a fresh JFS LVM2 volume a heavily used for month, filled with 4 million files reiserfs still outperforms JFS
10:34 < rxr> svn up in a t2-trunk in 28s vs. 1m22
10:34 < rxr> guess I have to try XFS, next ...
10:35 < rxr> the only annoying thing about reiser3 are thos journal commit storms where most I/O is nearly freezing on heavy I/O
10:35 < Ragnarin> yea
10:37 < rxr> Ragnarin: do you run "big" raid/lvm things?
10:37 < rxr> if yes, which FS do you prefer these days ?
10:39 < Ragnarin> no
10:39 < Ragnarin> but reiser
10:39 < Ragnarin> ;)
10:40 < rxr> heh
10:40 < Ragnarin> don't have big write load
10:40 < rxr> evertime I try the outher candidates out they have poorer performance even in the first hours of testing on a fres volume ...
10:40 < Ragnarin> yea
10:40 < Ragnarin> /dev/sda8 905367500K 258068020K 647299480K 29% /export
10:40 < Ragnarin> reiser
11:22 -!- mqueiros_ [n=mqueiros@c-217-70-70-61.bragatel.pt] has quit []
11:45 < mtr> moin
11:50 < rxr> moin moin mtr
12:10 < mtr> groff configure script runs a csh which in turn creates a /root/.history <-- evil in stage 1!
12:12 < rxr> ui
12:12 < mtr> (the host system is OpenSuse, not T2)
12:13 < rxr> (patch welcome :-)
12:13 < mtr> a. remove csh from build host
12:13 < mtr> b. patch groff
12:13 < rxr> patch
12:14 < mtr> yep, patch is on the way
12:19 < mtr> CIA ?
12:21 < rxr> died some days ago
12:30 -!- d-marc [n=d-marc@dslb-084-058-067-139.pools.arcor-ip.net] has joined #t2
12:31 < rxr> moin d-marc
12:31 < rxr> d-marc: long not seen - how's life ?
12:31 < d-marc> hello
12:31 < d-marc> I am ok, and you?
12:32 < rxr> d-marc: fine, too thanks
12:32 < d-marc> I just finished setting up the furniture last week, after I moved back to FFM
12:32 < rxr> how's your new job ?
12:32 < d-marc> pretty good
12:33 < d-marc> I dont have to use windows anymore, except for lotus notes ;)
12:33 < d-marc> now its AIX and solaris
12:34 < d-marc> how is T2 going?
12:36 < rxr> pretty good
12:36 < rxr> (IMHO)
12:37 < d-marc> hehe
12:37 < d-marc> last week tommi and I did our first steps towards the tntbuild server
12:38 < d-marc> maybe thats interesting for you too
12:38 < d-marc> It is a webapp that constantly checks if something has changed in SVN
12:38 < d-marc> and if so it starts a new compile
12:39 < d-marc> if it goes wrong people get email notification
12:50 < rxr> yeah - we use simillar stuff (just not a web app) for the "small" projects, such as exact-image
12:51 < rxr> however for T2 the compile checks take too long
12:51 < rxr> so we despite the new 8-core w/ 8GB RAM we have in-depth check cycles of a week ...
12:51 < rxr> (that is some bigger native builds and cross building the major combinations ...)
13:05 < d-marc> do you try to detect changes in SVN before you start a build, or do you just rebuild totally every week?
13:14 < rxr> of course we can only rebuild changed parts
13:15 < rxr> but changes on the build scripts, architecture config and quirks, libtool magic fixup etc. can potentially change every package build ...
13:16 < rxr> so a just changed package rebuild for all architecture / c-library combinations just takes an hour or so on a fast box
13:16 < rxr> but a rebuild (from ccache) on scripts and other config changes for the major combinations really takes loooong
14:04 < d-marc> how do you detect a change in SVN? do you do 'svn up' and parse the outout?
14:07 < rxr> for t2, as it does not depend on SVN and other VC was used in the past by timestamps of the dependencies
14:30 -!- Baldzius [n=Baldzius@87.198.192.226] has joined #t2
14:38 < d-marc> off to town
14:38 < d-marc> see you later!
14:38 -!- d-marc [n=d-marc@dslb-084-058-067-139.pools.arcor-ip.net] has quit ["KVIrc 3.2.4 Anomalies http://www.kvirc.net/"]
14:48 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has joined #t2
15:01 < rxr> hm - interting
15:01 < rxr> snowing in Berlin
15:09 < Baldzius> nice
15:45 < rxr> Baldzius: the peformance of the LVM volume is limitted a little
15:45 < rxr> as I do performance test on the second logical volume ...
15:45 < rxr> your virtual's container's i/o performance is probably lagging a little ...
15:46 < Baldzius> yeah, i noticed
15:46 < Baldzius> but it's ok
15:46 < rxr> yeah - I just do the testing to hopefully find a more performing ssetup than the current Reiser3 on LVM2 ...
15:47 < rxr> though the JFS figures did not look that good so far
15:47 * rxr now testing xfs
15:48 < rxr> Baldzius: btw. the container has some non-zero resource limit counters
15:48 < rxr> did you experience build problems due to ram shortage or so ? or are the builds all just fine ?
15:49 < Baldzius> everything is ok
15:49 < Baldzius> nothing major
15:50 < rxr> ok - great
15:54 < Baldzius> fyi: merged pcre
15:54 < Baldzius> and testing new php from trunk
15:55 < Baldzius> also security related
15:55 < Baldzius> should be fine but just want to test before commiting
16:01 -!- shevy [n=shevy@chello080108103172.36.11.univie.teleweb.at] has joined #t2
16:02 < shevy> can you specify non-conformant (to the FHS standard) build targets i.e. gcc-4.2.2 would resider under /pkg/gcc/4.2.2 ?
16:13 -!- CIA-41 [n=CIA@208.69.182.149] has joined #t2
17:25 < Baldzius> =[:609 (last $?=0)> set +o xtrace
17:25 < Baldzius> Creating file list and doing final adaptions ...
17:25 < Baldzius> Created file outside allowed top-level dir: tmp
17:25 < Baldzius> Created file outside allowed top-level dir: tmp
17:25 < Baldzius> --- BUILD ERROR ---
17:25 < Baldzius> removed `/var/adm/logs/2-gcc.log'
17:25 < Baldzius> rxr: 7.0 HEAD ^^^^^^
17:45 -!- LMJ_p [n=LMJ@laf31-2-82-224-107-105.fbx.proxad.net] has quit ["Ex-Chat"]
17:47 -!- yokoy [n=yokoy@e178077202.adsl.alicedsl.de] has quit [Read error: 104 (Connection reset by peer)]
17:47 -!- Baldzius [n=Baldzius@87.198.192.226] has quit ["Leaving"]
17:51 -!- yokoy [n=yokoy@e178077202.adsl.alicedsl.de] has joined #t2
20:44 -!- Netsplit kornbluth.freenode.net <-> irc.freenode.net quits: yokoy
20:45 -!- Netsplit kornbluth.freenode.net <-> irc.freenode.net quits: @ChanServ
20:46 -!- Netsplit over, joins: @ChanServ
20:46 -!- Netsplit over, joins: yokoy
20:47 -!- shevy [n=shevy@chello080108103172.36.11.univie.teleweb.at] has left #t2 ["I'll be back"]
22:29 -!- CIA-41 [n=CIA@208.69.182.149] has quit []
22:30 -!- Netsplit kornbluth.freenode.net <-> irc.freenode.net quits: @ChanServ
22:32 -!- Netsplit over, joins: @ChanServ
22:37 -!- CIA-12 [n=CIA@208.69.182.149] has joined #t2
22:38 -!- Netsplit kornbluth.freenode.net <-> irc.freenode.net quits: @ChanServ
22:39 -!- Netsplit kornbluth.freenode.net <-> irc.freenode.net quits: yokoy
22:40 -!- Netsplit over, joins: yokoy
22:43 -!- ChanServ [ChanServ@services.] has joined #t2
22:43 -!- ServerMode/#t2 [+o ChanServ] by irc.freenode.net
22:48 -!- Netsplit kornbluth.freenode.net <-> irc.freenode.net quits: @ChanServ
22:50 -!- Netsplit over, joins: @ChanServ
22:56 -!- ChanServ [ChanServ@services.] has quit [Shutting Down]
23:05 -!- ChanServ [ChanServ@services.] has joined #t2
23:05 -!- ServerMode/#t2 [+o ChanServ] by irc.freenode.net
--- Log closed Sun Nov 11 00:00:16 2007
05:07 -!- yokoy [n=yokoy@e178077202.adsl.alicedsl.de] has joined #t2
05:21 -!- yokoy_ [n=yokoy@e178075071.adsl.alicedsl.de] has quit [Read error: 110 (Connection timed out)]
08:13 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has quit ["There is intelligent life out there..."]
10:19 < rxr> re
10:19 < rxr> Ragnarin: which libtool build failure are you looking at ?
10:19 < Ragnarin> see the last few commits
10:19 < rxr> Ragnarin: the libtool hackery often is a new jungle, I think there is nothing like "not this dir"
10:19 < Ragnarin> wasn't actually libtools fault
10:19 < rxr> oh
10:19 < rxr> cia where are you ????
10:19 < Ragnarin> the package had a -L/usr/lib hardcoded
10:20 < rxr> finally:
10:20 < rxr> Error logs from default-7.0-stable-generic-sparc64-generic:
10:20 < rxr> 231 builds total, 231 completed fine, 0 with errors, 0 to go.
10:33 < rxr> hm - even on a fresh JFS LVM2 volume a heavily used for month, filled with 4 million files reiserfs still outperforms JFS
10:34 < rxr> svn up in a t2-trunk in 28s vs. 1m22
10:34 < rxr> guess I have to try XFS, next ...
10:35 < rxr> the only annoying thing about reiser3 are thos journal commit storms where most I/O is nearly freezing on heavy I/O
10:35 < Ragnarin> yea
10:37 < rxr> Ragnarin: do you run "big" raid/lvm things?
10:37 < rxr> if yes, which FS do you prefer these days ?
10:39 < Ragnarin> no
10:39 < Ragnarin> but reiser
10:39 < Ragnarin> ;)
10:40 < rxr> heh
10:40 < Ragnarin> don't have big write load
10:40 < rxr> evertime I try the outher candidates out they have poorer performance even in the first hours of testing on a fres volume ...
10:40 < Ragnarin> yea
10:40 < Ragnarin> /dev/sda8 905367500K 258068020K 647299480K 29% /export
10:40 < Ragnarin> reiser
11:22 -!- mqueiros_ [n=mqueiros@c-217-70-70-61.bragatel.pt] has quit []
11:45 < mtr> moin
11:50 < rxr> moin moin mtr
12:10 < mtr> groff configure script runs a csh which in turn creates a /root/.history <-- evil in stage 1!
12:12 < rxr> ui
12:12 < mtr> (the host system is OpenSuse, not T2)
12:13 < rxr> (patch welcome :-)
12:13 < mtr> a. remove csh from build host
12:13 < mtr> b. patch groff
12:13 < rxr> patch
12:14 < mtr> yep, patch is on the way
12:19 < mtr> CIA ?
12:21 < rxr> died some days ago
12:30 -!- d-marc [n=d-marc@dslb-084-058-067-139.pools.arcor-ip.net] has joined #t2
12:31 < rxr> moin d-marc
12:31 < rxr> d-marc: long not seen - how's life ?
12:31 < d-marc> hello
12:31 < d-marc> I am ok, and you?
12:32 < rxr> d-marc: fine, too thanks
12:32 < d-marc> I just finished setting up the furniture last week, after I moved back to FFM
12:32 < rxr> how's your new job ?
12:32 < d-marc> pretty good
12:33 < d-marc> I dont have to use windows anymore, except for lotus notes ;)
12:33 < d-marc> now its AIX and solaris
12:34 < d-marc> how is T2 going?
12:36 < rxr> pretty good
12:36 < rxr> (IMHO)
12:37 < d-marc> hehe
12:37 < d-marc> last week tommi and I did our first steps towards the tntbuild server
12:38 < d-marc> maybe thats interesting for you too
12:38 < d-marc> It is a webapp that constantly checks if something has changed in SVN
12:38 < d-marc> and if so it starts a new compile
12:39 < d-marc> if it goes wrong people get email notification
12:50 < rxr> yeah - we use simillar stuff (just not a web app) for the "small" projects, such as exact-image
12:51 < rxr> however for T2 the compile checks take too long
12:51 < rxr> so we despite the new 8-core w/ 8GB RAM we have in-depth check cycles of a week ...
12:51 < rxr> (that is some bigger native builds and cross building the major combinations ...)
13:05 < d-marc> do you try to detect changes in SVN before you start a build, or do you just rebuild totally every week?
13:14 < rxr> of course we can only rebuild changed parts
13:15 < rxr> but changes on the build scripts, architecture config and quirks, libtool magic fixup etc. can potentially change every package build ...
13:16 < rxr> so a just changed package rebuild for all architecture / c-library combinations just takes an hour or so on a fast box
13:16 < rxr> but a rebuild (from ccache) on scripts and other config changes for the major combinations really takes loooong
14:04 < d-marc> how do you detect a change in SVN? do you do 'svn up' and parse the outout?
14:07 < rxr> for t2, as it does not depend on SVN and other VC was used in the past by timestamps of the dependencies
14:30 -!- Baldzius [n=Baldzius@87.198.192.226] has joined #t2
14:38 < d-marc> off to town
14:38 < d-marc> see you later!
14:38 -!- d-marc [n=d-marc@dslb-084-058-067-139.pools.arcor-ip.net] has quit ["KVIrc 3.2.4 Anomalies http://www.kvirc.net/"]
14:48 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has joined #t2
15:01 < rxr> hm - interting
15:01 < rxr> snowing in Berlin
15:09 < Baldzius> nice
15:45 < rxr> Baldzius: the peformance of the LVM volume is limitted a little
15:45 < rxr> as I do performance test on the second logical volume ...
15:45 < rxr> your virtual's container's i/o performance is probably lagging a little ...
15:46 < Baldzius> yeah, i noticed
15:46 < Baldzius> but it's ok
15:46 < rxr> yeah - I just do the testing to hopefully find a more performing ssetup than the current Reiser3 on LVM2 ...
15:47 < rxr> though the JFS figures did not look that good so far
15:47 * rxr now testing xfs
15:48 < rxr> Baldzius: btw. the container has some non-zero resource limit counters
15:48 < rxr> did you experience build problems due to ram shortage or so ? or are the builds all just fine ?
15:49 < Baldzius> everything is ok
15:49 < Baldzius> nothing major
15:50 < rxr> ok - great
15:54 < Baldzius> fyi: merged pcre
15:54 < Baldzius> and testing new php from trunk
15:55 < Baldzius> also security related
15:55 < Baldzius> should be fine but just want to test before commiting
16:01 -!- shevy [n=shevy@chello080108103172.36.11.univie.teleweb.at] has joined #t2
16:02 < shevy> can you specify non-conformant (to the FHS standard) build targets i.e. gcc-4.2.2 would resider under /pkg/gcc/4.2.2 ?
16:13 -!- CIA-41 [n=CIA@208.69.182.149] has joined #t2
17:25 < Baldzius> =[:609 (last $?=0)> set +o xtrace
17:25 < Baldzius> Creating file list and doing final adaptions ...
17:25 < Baldzius> Created file outside allowed top-level dir: tmp
17:25 < Baldzius> Created file outside allowed top-level dir: tmp
17:25 < Baldzius> --- BUILD ERROR ---
17:25 < Baldzius> removed `/var/adm/logs/2-gcc.log'
17:25 < Baldzius> rxr: 7.0 HEAD ^^^^^^
17:45 -!- LMJ_p [n=LMJ@laf31-2-82-224-107-105.fbx.proxad.net] has quit ["Ex-Chat"]
17:47 -!- yokoy [n=yokoy@e178077202.adsl.alicedsl.de] has quit [Read error: 104 (Connection reset by peer)]
17:47 -!- Baldzius [n=Baldzius@87.198.192.226] has quit ["Leaving"]
17:51 -!- yokoy [n=yokoy@e178077202.adsl.alicedsl.de] has joined #t2
20:44 -!- Netsplit kornbluth.freenode.net <-> irc.freenode.net quits: yokoy
20:45 -!- Netsplit kornbluth.freenode.net <-> irc.freenode.net quits: @ChanServ
20:46 -!- Netsplit over, joins: @ChanServ
20:46 -!- Netsplit over, joins: yokoy
20:47 -!- shevy [n=shevy@chello080108103172.36.11.univie.teleweb.at] has left #t2 ["I'll be back"]
22:29 -!- CIA-41 [n=CIA@208.69.182.149] has quit []
22:30 -!- Netsplit kornbluth.freenode.net <-> irc.freenode.net quits: @ChanServ
22:32 -!- Netsplit over, joins: @ChanServ
22:37 -!- CIA-12 [n=CIA@208.69.182.149] has joined #t2
22:38 -!- Netsplit kornbluth.freenode.net <-> irc.freenode.net quits: @ChanServ
22:39 -!- Netsplit kornbluth.freenode.net <-> irc.freenode.net quits: yokoy
22:40 -!- Netsplit over, joins: yokoy
22:43 -!- ChanServ [ChanServ@services.] has joined #t2
22:43 -!- ServerMode/#t2 [+o ChanServ] by irc.freenode.net
22:48 -!- Netsplit kornbluth.freenode.net <-> irc.freenode.net quits: @ChanServ
22:50 -!- Netsplit over, joins: @ChanServ
22:56 -!- ChanServ [ChanServ@services.] has quit [Shutting Down]
23:05 -!- ChanServ [ChanServ@services.] has joined #t2
23:05 -!- ServerMode/#t2 [+o ChanServ] by irc.freenode.net
--- Log closed Sun Nov 11 00:00:16 2007