=== jb-home [n=jbailey@modemcable139.249-203-24.mc.videotron.ca] has joined #ubuntu-ports | ||
fabbione | Brad_: it's the same amount of fixes from the final kernel | 01:34 |
---|---|---|
fabbione | tmarble: no, i don't think so | 01:34 |
tmarble | thanks... FWIW we tried to install your new netboot image on an E4500 and had some troubles | 01:35 |
tmarble | first time we got Memory Address not Aligned | 01:36 |
tmarble | then next time we tried a WAN Boot: http://docs.sun.com/app/docs/doc/819-2394/6n4mg9dr0?q=%22boot+net%22&a=view | 01:36 |
tmarble | and got TFTP Error: Access violation, boot load failed | 01:37 |
tmarble | so I've suggested setting up a local tftpboot server for that machine | 01:37 |
tmarble | I still may hear from Ireland tomorrow and there is another T2000 that I might get access to | 01:38 |
tmarble | I'll let you know as soon as I hear anything | 01:39 |
fabbione | yeah don't sweat too much on non niagara hw | 01:40 |
tmarble | did you guys build the kernel with sun4v specific optimizations? | 01:41 |
tmarble | (or libs) | 01:41 |
fabbione | let me take some cofffee and i will try to answer | 01:42 |
tmarble | no worries | 01:42 |
fabbione | it's 1:40 am and i woke up because of the releas | 01:42 |
fabbione | +e | 01:42 |
tmarble | I have to go off line now anyway... | 01:42 |
tmarble | so please get some rest! | 01:42 |
=== fabbione has a release to do | ||
fabbione | will rest later | 01:43 |
tmarble | I'll try to get online early tomorrow and check with Dublin | 01:43 |
fabbione | i hope it will be released by the time you wake up | 01:43 |
fabbione | :) | 01:43 |
tmarble | ah | 01:43 |
fabbione | or that means i will be frying my brain in here | 01:43 |
tmarble | the truth is this new software on new hardware -- there will likely be more than one post release update | 01:44 |
tmarble | so don't sweat it! | 01:44 |
fabbione | there is a diff on what lands on CD and what's available for netboot/netinstall | 01:45 |
fabbione | CD are pressed | 01:45 |
fabbione | the latter isn't | 01:45 |
tmarble | understood... I'm sorry I can't give you any confirmations now... my hardware in Boston is unavailable -- the Ireland guys are sleeping... the third machine locally hasn't been authorized for reformatting and the guy in Colorado doesn't have any Niagara's in his lab (until July) | 01:47 |
fabbione | it's ok :) | 01:47 |
fabbione | thanks for the effort tho | 01:47 |
tmarble | doing what I can... | 01:48 |
tmarble | I have to go now.. good luck and I'll check in tomorrow | 01:48 |
fabbione | thanks | 01:48 |
=== Brad_ will do a full install with that image when I get home (probably about 2 hours from now) | ||
fabbione | hopefully by that time we will have an iso to test | 02:55 |
=== ajmitch__ [n=ajmitch@203.89.166.123] has joined #ubuntu-ports | ||
=== ajmitch__ [n=ajmitch@203.89.166.123] has joined #ubuntu-ports | ||
braddr | fabbione: if you have iso's you like tested on my t2k, I'd be happy to download, burn, and test | 06:13 |
=== ajmitch__ [n=ajmitch@203.89.166.123] has joined #ubuntu-ports | ||
fabbione | braddr: not yet ... we are getting there | 07:22 |
=== braddr nods. | ||
=== braddr boots up the box to make sure disk1 is nukeable. | ||
fabbione | you can still test netinstall in the meantime | 07:24 |
braddr | if you need another tester for it.. sure. I was assuming that it was a well covered case already | 07:25 |
fabbione | yeah but it's always better to do one more than one less | 07:26 |
braddr | true, see also the problems we found with my box and the kernel. | 07:26 |
fabbione | http://archive.ubuntu.com/ubuntu/dists/dapper-updates/main/installer-sparc/current/images/sparc64/netboot/2.6/boot.img | 07:26 |
braddr | looks good so far.. it's well into the first stage of the install, downloading additional components | 07:51 |
fabbione | nice | 07:51 |
=== braddr recalls you posting a testing matrix url ages ago. Is there anything in particular you need tested.. some dimension other than straight through boring install? | ||
fabbione | nah | 07:53 |
fabbione | that code is very common on all arches | 07:53 |
braddr | okey | 07:53 |
fabbione | i did test stuff like raid+lvm or just lvm | 07:53 |
braddr | right.. going the simple route.. use whole disk, just do it. | 07:55 |
fabbione | do whatever you feel like :) | 07:55 |
fabbione | that's how usually stuff breaks | 07:55 |
=== braddr wishes he had a faster net connection | ||
fabbione | http://cdimage.ubuntu.com/ubuntu-server/ports/daily/20060616.3/dapper-server-sparc.iso | 08:57 |
fabbione | there | 08:57 |
fabbione | start testing that one | 08:57 |
=== braddr ponders.. finish this install then download the .iso or abort and start over. How big is the image, a full 650meg? | ||
braddr | over an hour left on the first install's downloading left. | 08:59 |
fabbione | halt | 08:59 |
fabbione | wrong image :) | 08:59 |
fabbione | -rw-r--r-- 23306240 2006/06/16 08:55:07 dapper-server-sparc.iso | 08:59 |
fabbione | hmmm | 08:59 |
braddr | only 23 megs? Kinda tiny | 09:00 |
fabbione | something wrong | 09:00 |
fabbione | they are checking | 09:01 |
=== braddr nods, "I'll probably let this install finish first anyway." | ||
fabbione | wise | 09:12 |
=== fabbione starts to summon some greeks gods | ||
fabbione | it seems we just found a nice mkisofs bug | 09:17 |
braddr | the joys of release nights. | 09:20 |
=== fabbione wears his excorcist clothes | ||
braddr | useful when installing all those daemons. | 09:21 |
fabbione | "MOMMY MOMMY! there is a BSD daemon under my bed!" | 09:21 |
fabbione | rsync cdimage.ubuntu.com::cdimage/ubuntu-server/daily/20060616.3/dapper-server-sparc.iso | 09:56 |
fabbione | this one should be good | 09:56 |
fabbione | it's about 390M | 09:58 |
braddr | looks like another 15 minutes of downloading left on the first install.. but soon. | 10:01 |
braddr | though I probably shouldn't stay up that late | 10:01 |
fabbione | braddr: yeah well i can only test cdrom on 2 machines | 10:04 |
fabbione | if they work for me, they have to be ok for * | 10:04 |
fabbione | one test more will really be appreciated | 10:04 |
fabbione | but i can understand if you need sleep | 10:04 |
=== braddr checks to see how early his first meeting is | ||
braddr | 10 hrs from now.. so I can spare some more time. | 10:05 |
fabbione | GREAT! | 10:05 |
braddr | does it still have a second stage download across the net or are both parts done off cd? | 10:06 |
fabbione | this is -server cd | 10:06 |
fabbione | you get only stage one to install | 10:06 |
fabbione | + daemons on cd that are not installed by default | 10:06 |
braddr | hrm.. I don't actually have a blank cdr, but I do have dvd blanks. | 10:06 |
fabbione | it would do | 10:07 |
fabbione | makes no diff | 10:07 |
fabbione | it's an iso | 10:07 |
braddr | I figured as much, just a waste. :) | 10:07 |
fabbione | well ude a dvd-rw | 10:08 |
braddr | of, uh.. what.. 20 cents? :P | 10:08 |
fabbione | use even | 10:08 |
braddr | they're dvd+r's which are multi-write capable | 10:09 |
fabbione | cd boots here | 10:10 |
braddr | always a plus | 10:11 |
fabbione | it doesn't install... | 10:12 |
fabbione | but start downloading it | 10:12 |
fabbione | rsync will be fast | 10:12 |
braddr | not a good sign. | 10:12 |
fabbione | i know what the problem is | 10:12 |
fabbione | yeah fixing it now | 10:13 |
fabbione | don't stop downloading it | 10:13 |
braddr | haven't started it yet.. the first install's download phase just finished. | 10:14 |
braddr | but I'm starting it now | 10:14 |
fabbione | cdimage.ubuntu.com::cdimage/ubuntu-server/daily/20060616.5/dapper-server-sparc.iso | 10:46 |
fabbione | new image | 10:46 |
fabbione | this one should be good | 10:46 |
=== braddr waits for the first rsync to finish. | ||
fabbione | yeah the diff is like 1MB | 10:48 |
fabbione | it will be fast in the second shot | 10:48 |
fabbione | 6747622 | 10:48 |
fabbione | sorry a bit more .. 6MB | 10:48 |
braddr | whatever. :) | 10:49 |
fabbione | yeah | 10:49 |
braddr | how long should the 'cleaning up' step of the 'select and install software' phase? It's been sitting there for several minutes now. | 11:33 |
braddr | oh sure.. it just moved on. :) | 11:34 |
fabbione | yeah it's long and slow | 11:36 |
braddr | whee.. finally finished that first install.. went well. | 11:43 |
fabbione | cool | 11:43 |
fabbione | i am at the second cd install here on the t2k | 11:44 |
fabbione | with lamp-server | 11:44 |
braddr | the 2nd rsync is started. | 11:44 |
fabbione | if you are booting from cd, please do a lamp-server-expert | 11:44 |
fabbione | it will bomb you with questions | 11:44 |
fabbione | but it's a test case i didn't do | 11:44 |
braddr | sure. | 11:44 |
braddr | care to double check the md5sum for the iso: 82da246064785adb7b87e5aa0cb5764c | 11:46 |
braddr | ah.. nevermind, I see the md5sums file | 11:46 |
fabbione | ;) | 11:49 |
braddr | either there's a difference between cd and dvd .iso formats, or this windows burning software sucks | 11:56 |
fabbione | i'd say the latter :) | 11:58 |
braddr | boots up this vmware linux install to see if it can write it | 12:01 |
braddr | argh.. it's expired. <sigh> | 12:01 |
fabbione | meh | 12:04 |
fabbione | well it's booting here | 12:04 |
=== braddr ponders.. loopback mount of the .iso file on one of my other machines, nfs export it, then use netboot and nfs? :P | ||
fabbione | so whatever reason it doesn't, it's local | 12:04 |
braddr | I haven't gotten the image written yet.. the windows software is requiring a cd and not a dvd blank | 12:05 |
fabbione | bah | 12:06 |
fabbione | grwoisofs :) | 12:06 |
fabbione | growisofs even | 12:06 |
=== braddr is considering pretty strongly opting for sleep instead of spending another couple hours on this install. | ||
braddr | as much as I'd like to have completed another test case for you. | 12:07 |
fabbione | go to sleep | 12:07 |
fabbione | don't worry | 12:08 |
fabbione | we will manage | 12:08 |
braddr | I can/will snag another stack of cdrw's on the way home from work tomorrow | 12:08 |
fabbione | we will be released by that time | 12:09 |
fabbione | :) | 12:09 |
braddr | doesn't totally obviate the need for that case to be tested. :) | 12:09 |
fabbione | well in theory it works | 12:10 |
fabbione | if it doesn.. we will just say.. Whoops ;) | 12:10 |
braddr | .. and then put out an update. :) But first you gotta find out if theory and practice are on speaking terms. | 12:11 |
fabbione | ehhe | 12:12 |
=== braddr just took a look at the commit with the hack for that bug.. rather simple | ||
fabbione | http://releases.ubuntu.com/dapper/ | 01:31 |
fabbione | there | 01:31 |
=== tmarble [n=tmarble@192.18.101.5] has joined #ubuntu-ports | ||
fabbione | https://lists.ubuntu.com/archives/ubuntu-announce/2006-June/000087.html | 02:18 |
fabbione | go nuts | 02:18 |
fabbione | i am off to get some sleep | 02:18 |
tmarble | buon nuitto | 02:18 |
fabbione | some mirrors are still syncing | 02:18 |
fabbione | but they will get there eventually :) | 02:18 |
tmarble | congratulations! | 02:19 |
fabbione | thanks | 02:19 |
tmarble | after you have had some sleep I'll ask you to elaborate on "Niagara optimizations for gcc, glibc and kernel available by default" | 02:20 |
tmarble | good work! get some rest! | 02:21 |
fabbione | it's very simple | 02:21 |
fabbione | gcc-4.0 had -mcpu=niagara | 02:21 |
fabbione | glibc has optmized versions for niagara and they are installed by default | 02:22 |
fabbione | kernel is sort of "support for Niagara" | 02:22 |
fabbione | edgy will be even faster than dapper | 02:22 |
fabbione | glibc are compiled with gcc-4.1 and -mcpu=niagara | 02:22 |
fabbione | something i was not comfrotable to do in dapper | 02:22 |
tmarble | when you say faster... how do you know? | 02:22 |
=== tmarble asks the performance benchmarking analyst ;-) | ||
fabbione | simply because you use optimized code for that specific cpu | 02:23 |
fabbione | instead of some generic one | 02:23 |
fabbione | for example niagara has a better handling for memcpy | 02:24 |
fabbione | but if you don't tell userland how to use it | 02:24 |
fabbione | it will use the generic implementation | 02:24 |
fabbione | so the optimized glibc will use the better memcpy implementation for that cpu | 02:24 |
tmarble | so if I install Ubuntu on a non-niagara (older) system do I need to be careful to choose a non-optimized kernel and glibc? | 02:24 |
fabbione | tmarble: no, it's all done transparently | 02:25 |
fabbione | you don't need to worry about anything | 02:25 |
fabbione | if it is possible we use optimizations, otherwise no | 02:25 |
fabbione | you can see using ldd /bin/ls for example | 02:25 |
fabbione | if you are using niagara you get something like /lib/v9v/ | 02:25 |
tmarble | I won't bug you with more questions now (but I'd love to understand how the magic works!) | 02:25 |
fabbione | otherwise it's just /lib | 02:25 |
fabbione | there are also ultra3 optimizations (v9b) | 02:26 |
fabbione | it's easy | 02:26 |
fabbione | the dynamic lynker checks for cpu capabilities | 02:26 |
fabbione | if it matches and there are libs that can explicitly exploit that capability, they will be preferred | 02:26 |
tmarble | wow.. that's cool... so could we use that to transparently support 32-bit and 64-bit packages on amd64 (for example) | 02:27 |
fabbione | that's called multiarch | 02:27 |
fabbione | it's slightly different target but it will happen for edgy+1 | 02:28 |
fabbione | running 32bit apps on a 64bit kernel/userland | 02:28 |
fabbione | or a mixed userland | 02:28 |
fabbione | on sparc userland is mainly at 32 bit | 02:28 |
fabbione | but there is enough to build 64bit | 02:29 |
fabbione | tho the latter should be used carefully | 02:29 |
fabbione | since it is generally slower | 02:29 |
tmarble | right.... I will need to learn more about this co-existance (i've heard a lot about multiarch) | 02:29 |
fabbione | tmarble: you want to talk to Mithrandir and Keybuk on #ubuntu-devel | 02:30 |
fabbione | since they have been working on it already | 02:30 |
tmarble | cool | 02:30 |
=== fabbione yawns | ||
fabbione | this was a good sleep | 04:32 |
tmarble | wow, that was short! | 04:46 |
fabbione | well it's 5pm | 04:49 |
fabbione | almost | 04:49 |
fabbione | i don't want to spend the day sleeping :) | 04:49 |
fabbione | already one (duplicate) bug :) | 04:54 |
braddr | congrats fabbione | 07:14 |
braddr | (and crew) | 07:15 |
fabbione | braddr: thanks | 07:16 |
braddr | any reason the release announcement didn't reference any sun url's or thank them for their support? goodwill building stuff. | 07:18 |
fabbione | braddr: yes.. my t2000 has only 6 cores and 8G of RAm... they promised a desktop and i didn't get it :P | 07:19 |
fabbione | so they don't deserve space in the announcement ;) | 07:19 |
braddr | hahahhaha.. so lack of goodwill. :P | 07:19 |
fabbione | i am just kidding of course | 07:19 |
fabbione | it's just the standard template | 07:19 |
tmarble | congratulations everyone: http://blogs.sun.com/roller/page/jonathan?entry=ubuntu_on_niagara_and_platinum | 11:37 |
tmarble | and, i | 11:37 |
tmarble | have my own T2000 (for a couple weeks anyway) and I'm installing now... | 11:37 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!