Re: [osol-discuss] SchilliX-0.7.0 ready for testing
On 07/24/10 08:23 PM, Shawn Walker wrote: On 07/24/10 04:17 PM, Ian Collins wrote: On 07/25/10 10:27 AM, Shawn Walker wrote: On 07/24/10 06:37 AM, Moinak Ghosh wrote: ... To stir the pot here, since we are discussing a "Community Distro" as opposed to a SUN/Oracle distro, IPS when used remotely from halfway across the world has large performance issues. For example in Bangalore I personally know no one outside the SUN India office who have successfully updated packages on their OpenSolaris installations let alone do an image-update in a sane amount of time. In comparison things like YUM or Apt-Get are reasonable even over 512Kbps. Having said that I am not going to make vague statements. I will do my own testing over a 2Mbps broadband link and post measurements. The version of pkg(5) that will be part of b144 should deliver somewhere around a 20% or greater performance improvement in transport performance when used with a properly configured web and/or depot server. If we ever get to see b > 134! The source is there, and you can build it you know ;) Alternatively, you could use the bits someone made available on genunix.org for b142: http://genunix.org/dist/richlowe/README.txt http://genunix.org/dist/richlowe/ Hi, Shawn! Do you know about any tutorial / white paper / something explaining how to start from the sources, and create a repo such as Rich Lowe did with snv_142? I wrote something in English, and posted, and something in Spanish for our community site about how to upgrade from snv_134 to onnv_142, but don't know how to start from sources and end with such a repo. I understand Rich did a great job with this, and want to reproduce it locally. Any help will be appreciated. Thanks, and best regards, HeCSa. Cheers, -Shawn ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SchilliX-0.7.0 ready for testing
>This is not an IPS issue; this is a distribution issue. Remember that >all of the bits on pkg.opensolaris.org have been redistributable for a >long time. That means that if someone wanted a mirror in say, Finland, >they could have set one up. For that matter, one could have been setup >on genunix.org a while ago. IPS tries to be a network pkg management system. It needs to be simple and easy to be used by folks to create simple mirrors and maintain them. IPS stays at the heart of this process. You need to measure how many such mirrors are available worldwide and try to understand why there are few, many etc. Then, there should be as well a connection between installer, pkg management and all these repositories/mirrors worldwide. The OSOL installer, IPS should keep a list of all hot mirrors available and automatically select the one close to you. Your TIMEZONE -> auto select a close repo -> validate -> if ok install from there if not fall back to another one Otherwise again all these things are disconnected and users are left in dark ! >Fixes went in builds after b134 that reduced memory usage by as much as 60%. >At the moment (after the fixes noted above), memory usage remains >primarily a function of the amount of package data the client has to >process. Considering that the /dev repository contains nearly 70,000 >unique package versions, I'd say it's pretty good at the moment. Thanks. Looks good. What comparative analysis have you done IPS vs SVR4 pkg management regarding: memory usage, cpu or net usage , resource consumption in general ? Do you have any open documents where we can see how good IPS is versus SVR4 in terms of resource utilisation ? As well what means in terms of CPU/MEM/Net to install 2000 pkg with IPS versus SVR4 ? Thanks for all updates. stefan -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] [distribution-discuss] SchilliX-0.7.0 ready for testing
On 07/24/10 11:09 AM, Martin Bochnig wrote: 2010/7/24 "C. Bergström": Joerg Schilling wrote: "C. Bergström" wrote: Did you pull the libc work from Garrett and drop that on there? No, I just created a halfway clean base for starting with the emancipation work. Which compiler did you use to build onnv-gate? OS-12.1 That's not a compiler version. I know gcc doesn't boot, but ping me off list or on irc if you'd like to help test a compiler to replace SS That´s a vague statement. Last time I did this gcc compiled 64bit kernels did boot (sun4u, sun4v, amd64). The 32bit x86 kernel alone was unbootable (and who wants that nowadays, except maybe for embedded use). Some people like me want it, because of a lot of notebooks, and netbooks too. I'm running snv_134 in two 32 bits machines, and one 64 bits one! Don't forget us! Thanks, and best regards, HeCSa. ___ distribution-discuss mailing list distribution-disc...@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/distribution-discuss ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Community distro
On 07/24/10 04:30 PM, joerg.schill...@fokus.fraunhofer.de wrote: Alan Coopersmith wrote: joerg.schill...@fokus.fraunhofer.de wrote: Alan Coopersmith wrote: Can Oracle Engineering provide someone like Joerg what he needs to update his external SchilliX distro to implement ON snv_145/IPS 134? I'm guessing you're talking to me? I have no idea what Joerg needs, but we've already provided the community with everything they need to build current ON& IPS bits, as evidenced by multiple external community members doing so. You do need a recent build of IPS to build ON builds 136& later, and X builds 144& later, but that's fully available from the pkg-gate sources on the opensolaris hg repos. Do you have a IPS SVr4 package or do you have a IPS source that compiles on a machine without IPS? The IPS source compiles on machines without IPS. Is there a separate IPS source? hg clone ssh://a...@hg.opensolaris.org/hg/pkg/gate The IPS source that is inside b138 did not completely compile as it expects definition files to be in the host OS that are missing on SCXE b130 Yes, you will need a version of ON newer than what you'll find in SXCE b130. pkg(5) isn't suited for use with SXCE. The source for ON builds > 130 is available, and you can build it. If you don't want to build it yourself, feel free to use what Rich Lowe has provided at genunix.org (b142): http://genunix.org/dist/richlowe/ To use the above, you'll need to install b134a of the OpenSolaris distribution and then use the instructions in the README at the link above. Cheers, -Shawn ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Community distro
joerg.schill...@fokus.fraunhofer.de wrote: > Alan Coopersmith wrote: >> The IPS source compiles on machines without IPS. > > Is there a separate IPS source? The IPS source is hosted in the /hg/pkg/gate repository on src.opensolaris.org, as described on the IPS project web page: http://hub.opensolaris.org/bin/view/Project+pkg/WebHome > The IPS source that is inside b138 did not completely compile as it expects > definition files to be in the host OS that are missing on SCXE b130 If you're referring to ON b138, it uses IPS to build packages, but does not contain IPS itself, just as it uses compilers to build binaries, but does not contain Sun Studio or gcc itself. -- -Alan Coopersmith-alan.coopersm...@oracle.com Oracle Solaris Platform Engineering: X Window System ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SchilliX-0.7.0 ready for testing
On 07/25/10 11:23 AM, Shawn Walker wrote: On 07/24/10 04:17 PM, Ian Collins wrote: On 07/25/10 10:27 AM, Shawn Walker wrote: The version of pkg(5) that will be part of b144 should deliver somewhere around a 20% or greater performance improvement in transport performance when used with a properly configured web and/or depot server. If we ever get to see b > 134! The source is there, and you can build it you know ;) The source is also here, and I'm getting close to the point where I will! Once upon a time in a land far away I worked for a small company who chose to base a product on their own Linux distro. I warned them it would all end in tears, but they went ahead anyway. The support and testing costs killed it in the end. That's why I shy away from building my own OS! Alternatively, you could use the bits someone made available on genunix.org for b142: http://genunix.org/dist/richlowe/README.txt http://genunix.org/dist/richlowe/ I've seen that and it's a job very well done. -- Ian. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SchilliX-0.7.0 ready for testing
On 07/24/10 04:25 PM, joerg.schill...@fokus.fraunhofer.de wrote: Shawn Walker wrote: It isn't mean to be in SXCE, so wasn't put there. I think you misunderstand its purpose. Maybe IPS was not developed for integration, then I could understand why it did not appear on SXCE. IPS will not be used by e.g. swiss banks the way it currently works. They do not have internet on their servers. They have additional security components in their caverns in the mountains, they call these security components machine guns In order to get an OS on their servers, they like to have granted reproducability without network access. This worked with SXCE but this does not work with indina. You don't need network access to install packages using pkg(5). While it's true that in b134 you had to at least start a depot server, you could so on the local system itself. In builds 142+, you can actually configure the client to use a package repository on a filesystem directly so a depot server isn't required. The on-disk format for repositories is intended to be formalised soon (and this is no secret -- just watch pkg-discuss) and an on-disk archive format delivered. Cheers, -Shawn ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Community distro
Alan Coopersmith wrote: > joerg.schill...@fokus.fraunhofer.de wrote: > > Alan Coopersmith wrote: > > > >>> Can Oracle Engineering provide someone like Joerg what he needs to update > >>> his external SchilliX distro to implement ON snv_145/IPS 134? > >> I'm guessing you're talking to me? I have no idea what Joerg needs, but > >> we've > >> already provided the community with everything they need to build current > >> ON & > >> IPS bits, as evidenced by multiple external community members doing so. > >> You > >> do need a recent build of IPS to build ON builds 136 & later, and X builds > >> 144 & later, but that's fully available from the pkg-gate sources on the > >> opensolaris hg repos. > > > > Do you have a IPS SVr4 package or do you have a IPS source that compiles on > > a > > machine without IPS? > > The IPS source compiles on machines without IPS. Is there a separate IPS source? The IPS source that is inside b138 did not completely compile as it expects definition files to be in the host OS that are missing on SCXE b130 Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Community distro
joerg.schill...@fokus.fraunhofer.de wrote: > Alan Coopersmith wrote: > >>> Can Oracle Engineering provide someone like Joerg what he needs to update >>> his external SchilliX distro to implement ON snv_145/IPS 134? >> I'm guessing you're talking to me? I have no idea what Joerg needs, but >> we've >> already provided the community with everything they need to build current ON >> & >> IPS bits, as evidenced by multiple external community members doing so. You >> do need a recent build of IPS to build ON builds 136 & later, and X builds >> 144 & later, but that's fully available from the pkg-gate sources on the >> opensolaris hg repos. > > Do you have a IPS SVr4 package or do you have a IPS source that compiles on a > machine without IPS? The IPS source compiles on machines without IPS. -- -Alan Coopersmith-alan.coopersm...@oracle.com Oracle Solaris Platform Engineering: X Window System ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SchilliX-0.7.0 ready for testing
Shawn Walker wrote: > It isn't mean to be in SXCE, so wasn't put there. I think you > misunderstand its purpose. Maybe IPS was not developed for integration, then I could understand why it did not appear on SXCE. IPS will not be used by e.g. swiss banks the way it currently works. They do not have internet on their servers. They have additional security components in their caverns in the mountains, they call these security components machine guns In order to get an OS on their servers, they like to have granted reproducability without network access. This worked with SXCE but this does not work with indina. People who installed a bigger bunch of machines did not get the same results for all of them with IPS and called it "Internet Package Lottery". But as mentioned before, I don't believe that it is time to talk about packaging before we have a 100% OSS OS ready for deployment. Let us make the decision about packaging when it is time for such a decision and let each distro make the decision on which basic packaging to use by it's own. Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SchilliX-0.7.0 ready for testing
On 07/24/10 04:17 PM, Ian Collins wrote: On 07/25/10 10:27 AM, Shawn Walker wrote: On 07/24/10 06:37 AM, Moinak Ghosh wrote: ... To stir the pot here, since we are discussing a "Community Distro" as opposed to a SUN/Oracle distro, IPS when used remotely from halfway across the world has large performance issues. For example in Bangalore I personally know no one outside the SUN India office who have successfully updated packages on their OpenSolaris installations let alone do an image-update in a sane amount of time. In comparison things like YUM or Apt-Get are reasonable even over 512Kbps. Having said that I am not going to make vague statements. I will do my own testing over a 2Mbps broadband link and post measurements. The version of pkg(5) that will be part of b144 should deliver somewhere around a 20% or greater performance improvement in transport performance when used with a properly configured web and/or depot server. If we ever get to see b > 134! The source is there, and you can build it you know ;) Alternatively, you could use the bits someone made available on genunix.org for b142: http://genunix.org/dist/richlowe/README.txt http://genunix.org/dist/richlowe/ Cheers, -Shawn ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SchilliX-0.7.0 ready for testing
On 07/25/10 10:27 AM, Shawn Walker wrote: On 07/24/10 06:37 AM, Moinak Ghosh wrote: ... To stir the pot here, since we are discussing a "Community Distro" as opposed to a SUN/Oracle distro, IPS when used remotely from halfway across the world has large performance issues. For example in Bangalore I personally know no one outside the SUN India office who have successfully updated packages on their OpenSolaris installations let alone do an image-update in a sane amount of time. In comparison things like YUM or Apt-Get are reasonable even over 512Kbps. Having said that I am not going to make vague statements. I will do my own testing over a 2Mbps broadband link and post measurements. The version of pkg(5) that will be part of b144 should deliver somewhere around a 20% or greater performance improvement in transport performance when used with a properly configured web and/or depot server. If we ever get to see b > 134! -- Ian. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] b134 panics on bootup after install
I found it. Hit bug 6932552, had to disable the graphics boot from grub. Works great. Thanks -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] b134 panics on bootup after install
> I have installed b134 on a clevo laptop. The live cd > booted just fine and the install went on without a > problem. After the install when I reboot the system > comes up with the initial grub screen just after the > OS is selected the screen goes black and the system > reboots. Did you try tz boot in text mode with the kernel debugger enabled? What kind of panic do you get ? Edit the grub menu entry, remove the splashimage, foreground and background lines; remove the ",console=graphics" option from the kernel$ line, and append options " -kv" at the end of the kernel$ line. Boot using the modified entry. -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] b134 panics on bootup after install
I have installed b134 on a clevo laptop. The live cd booted just fine and the install went on without a problem. After the install when I reboot the system comes up with the initial grub screen just after the OS is selected the screen goes black and the system reboots. I thought I had seen this posted before but was unable to find anything about it in the archives. Anyone have any Ideas? Thanks -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SchilliX-0.7.0 ready for testing
On 07/24/10 07:06 AM, joerg.schill...@fokus.fraunhofer.de wrote: Shawn Walker wrote: On 07/23/10 04:41 PM, Ken Gunderson wrote: Why? IPS was shoved down the community's throat in a heavy handed and decidedly not FOSS manner. Sorry, but that's simply not true. The pkg(5) project has been one of the few projects that is actually very open. It was the first to use defect.opensolaris.org for bugtracking, it's licensed under the CDDL, and at the moment any contributor (even external ones) can get commit access upon approval by the project team members. The basic OSS rules are: - release early - release often The gate is public, so we release essentially every time there's a putback. If you mean a tarball, no, because this is an integrated distribution project. - port to many platforms. It runs on GNU/Linux, BSD, (Open)Solaris, and AIX (reportedly). Although I should note that the primary focus and drive for the project is a well integrated packaging system that meets the needs of the OS distribution it is integrated with. But there was no visibility and "testability" on SXCE and for this reason, it missed most users I know. Also not true. There are definitely individuals that use it on SXCE. Although I should note that's only for the depot server and user images. It is not intended to manage packages on a Solaris 10 system's '/' filesystem. For now, IPS is used by one OpenSolaris distro only and this distro published the last binary release nearly one year ago. That's also not true. b134a was published in March -- that's nowhere close to a year. Writing and publishing successful OSS is not only producing code. You need to convince the users also with facts. The best way this can be done is by making the software available everywhere (and creators of distributions have a big change with getting new users by including software), but IPS did not appear on SXCE... It isn't mean to be in SXCE, so wasn't put there. I think you misunderstand its purpose. Cheers, -Shawn ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SchilliX-0.7.0 ready for testing
On 07/24/10 10:08 AM, Stefan Parvu wrote: "> The only problem I can see in regards to how IPS was introduced that perhaps more explanation to the community at the project inception about why IPS and not something else would do good... a little bit more open dialog with the community before you start coding on why IPS, what are its design goals, etc. would help here." In my opinion, we will always have some gaps here, how open this process will be. Sun/Oracle will not ask the community unless they engage the community in the building process. In my view RedHat is a very well player which knows how to extract the goodies from Fedora without getting burned maintaining it ;) We need to learn to do the same. Otherwise we will live in hope. Regarding IPS and future we need to think and co-work this with vendor and find a common way together. IPS has still long way until it reaches some sort of stable status. From Finland some feedback I got from users regarding IPS (our internal FIOSUG, and other folks regarding IPS status found on 2009.06): - not a download format for a package. Meaning somebody would like to simple download emacs.pkg and manuall install it ! Cant do that now One is in development (right now). - repository cloning. Still hairy and poor documentation http://hub.opensolaris.org/bin/view/Project+pkg/Mirroring Also see pkgrecv(1). - mirrors: slow links to fetch packages from US. Who cares how well and smart IPS is if it takes forever to have some pkgs install in your machine. Ubuntu has a nice mirror list where pkgs are fetched within minutes. I can confirm as well this. Its a pain sometimes to download something from main repo. I live in Finland and have a 8Mbps line. This is not an IPS issue; this is a distribution issue. Remember that all of the bits on pkg.opensolaris.org have been redistributable for a long time. That means that if someone wanted a mirror in say, Finland, they could have set one up. For that matter, one could have been setup on genunix.org a while ago. - IPS resource consumption. Big and fat process when installing some basic pkgs. Dont have facts here but I have seen it too. I had the impression some work went in to get this fixed and improve things. Fixes went in builds after b134 that reduced memory usage by as much as 60%. At the moment (after the fixes noted above), memory usage remains primarily a function of the amount of package data the client has to process. Considering that the /dev repository contains nearly 70,000 unique package versions, I'd say it's pretty good at the moment. Cheers, -Shawn ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SchilliX-0.7.0 ready for testing
On 07/24/10 06:37 AM, Moinak Ghosh wrote: ... To stir the pot here, since we are discussing a "Community Distro" as opposed to a SUN/Oracle distro, IPS when used remotely from halfway across the world has large performance issues. For example in Bangalore I personally know no one outside the SUN India office who have successfully updated packages on their OpenSolaris installations let alone do an image-update in a sane amount of time. In comparison things like YUM or Apt-Get are reasonable even over 512Kbps. Having said that I am not going to make vague statements. I will do my own testing over a 2Mbps broadband link and post measurements. The version of pkg(5) that will be part of b144 should deliver somewhere around a 20% or greater performance improvement in transport performance when used with a properly configured web and/or depot server. Cheers, -Shawn ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SchilliX-0.7.0 ready for testing
Stefan Parvu wrote: > > Star-1.5.1 is also available from Blastwave. > > The star on SchilliX has one new feature: > > > > 1) create a meta data only archive with star -c -dump -meta . > arch > > 2) extraxt the archive uwing star -xp -xmeta -force-hole < arch > >in order to create all plain files as 100% empty holes > > New is the combination "-xmeta -force-hole". Such a tree does nse nearly no > > space on disk and it is sufficient as a reference for a wget -mirror > > call. > > > > sweet. star should have been long time inside Solaris. > I hope to see star first class citizen in future OSOL. I'll see whether I am able to create a set of makefiles for the Snorcle Build system in ON. Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SchilliX-0.7.0 ready for testing
On 07/25/10 02:57 AM, Ken Gunderson wrote: To all those who seemed to have been confused by my reply to Ian's post please note the following: 1) My comments were intended as a response to Ian's assertion of the 'inevitability of IPS'. The inevitability of IPS I was alluding to is Oracle's Solaris. My main reason for raising the subject was Jörg's comment about having to use an old Xorg. When a project has one contributor doing 99% of the work, not following their packaging scheme is a Quixotic gesture that will doom a distribution that lacks significant resources. Someone has to build packages. Nexenta's call to use Debian/RPM was a sound business decision. They can concentrate on adding value at the core while others supply the essential, but non-core packages. 2) The example I provided, Nexenta, is pretty strong evidence that incorporation of IPS is not requisite for a successful Open Solars derivative. True, but as I said, it's strong evidence for using a widely supported packaging scheme. 3) My comments did not even touch vaguely upon any issues of technical merit of package systems. Neither did mine, I am posting with my manager hat on! 4) I was here when Indiana was released. My recollection is that there was quite a bit of uproar about IPS, among other things, specifically in the context of the manner in which Sun played their trump card. Others are certainly free to interpret those events differently. There was, but it was largely piss and wind; no one put up an alternative. I disliked it at the time and said so, but now I have belatedly adopted it, I won't be going back (even if I could). 5) Funny that I'm accused of stirring the pot when the fact of the matter is that Ian used the ruse of a pat on the back as a thinly veiled cover for effectively telling Joerg that he'd better toe the Oracle IPS line. At least that was my read. Hence my response. Your mileage may vary. No, I didn't, there was no ruse I respect the work Jörg has done. I was expressing my view that if Jörg's distribution is to succeed, it needs to be able to draw on existing components from other sources. The OpenSolaris "community" outside of Oracle is too small to do any of devising, implementing or supporting yet another packaging scheme -- Ian. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Community distro
On Sat, Jul 24, 2010 at 12:07 PM, Joerg Schilling wrote: > Alan Coopersmith wrote: > >> > Can Oracle Engineering provide someone like Joerg what he needs to update >> > his external SchilliX distro to implement ON snv_145/IPS 134? >> >> I'm guessing you're talking to me? I have no idea what Joerg needs, but >> we've >> already provided the community with everything they need to build current ON >> & >> IPS bits, as evidenced by multiple external community members doing so. You >> do need a recent build of IPS to build ON builds 136 & later, and X builds >> 144 & later, but that's fully available from the pkg-gate sources on the >> opensolaris hg repos. > > Do you have a IPS SVr4 package or do you have a IPS source that compiles on a > machine without IPS? > > Jörg Some time ago (about the B130 era...) I'm pretty sure it built SVr4 packages. An existing installation is not required to build it. $ hg clone http://hg.genunix.org/pkg-gate.hg/ $ cd pkg-gate.hg/src $ export PATH=/opt/sunstudio12.1/bin:$PATH : fix a bunch of .po files with mismatched newlines : http://cr.opensolaris.org/~mgerdts/pkg-po.patch.gz $ dmake ... /usr/bin/python2.6 setup.py build running build running build_py running build_ext $ echo $? 0 $ uname -sr Linux 2.6.34-12-default It is intended to be very cross platform and is the underlying update mechanism in Oracle products other than OpenSolaris. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Community distro
--- On Sat, 7/24/10, Alan Coopersmith wrote: > Ken Mays wrote: > > Hi Alan, > > > > I'm taking a stab at this but I heard Joerg mentioned > an issue with moving beyond ON snv_130. Then, you brought up > some things about not being able to build ON 135+, X b143+, > GNOME/JDS, and a few things without IPS being a part of the > distro. So, IPS is definitely on the menu. > > > > Can Oracle Engineering provide someone like Joerg what > he needs to update his external SchilliX distro to implement > ON snv_145/IPS 134? > > I'm guessing you're talking to me? I have > no idea what Joerg needs, but we've > already provided the community with everything they need to > build current ON & > IPS bits, as evidenced by multiple external community > members doing so. You > do need a recent build of IPS to build ON builds 136 & > later, and X builds > 144 & later, but that's fully available from the > pkg-gate sources on the > opensolaris hg repos. > > -- > -Alan Coopersmith- > alan.coopersm...@oracle.com > Oracle Solaris Platform > Engineering: X Window System Well, I agree. Someone can still obtain the latest ON source and work with Sun Studio 12.1 to build it. Mostly have to go through the motions and burn a few candles. ~ Ken ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SchilliX-0.7.0 ready for testing
"> The only problem I can see in regards to how IPS was introduced that perhaps > more explanation to the community at the project inception about why IPS and > not something else would do good... a little bit more open dialog with the > community before you start coding on why IPS, what are its design goals, > etc. would help here." In my opinion, we will always have some gaps here, how open this process will be. Sun/Oracle will not ask the community unless they engage the community in the building process. In my view RedHat is a very well player which knows how to extract the goodies from Fedora without getting burned maintaining it ;) We need to learn to do the same. Otherwise we will live in hope. Regarding IPS and future we need to think and co-work this with vendor and find a common way together. IPS has still long way until it reaches some sort of stable status. From Finland some feedback I got from users regarding IPS (our internal FIOSUG, and other folks regarding IPS status found on 2009.06): - not a download format for a package. Meaning somebody would like to simple download emacs.pkg and manuall install it ! Cant do that now - repository cloning. Still hairy and poor documentation - mirrors: slow links to fetch packages from US. Who cares how well and smart IPS is if it takes forever to have some pkgs install in your machine. Ubuntu has a nice mirror list where pkgs are fetched within minutes. I can confirm as well this. Its a pain sometimes to download something from main repo. I live in Finland and have a 8Mbps line. - IPS resource consumption. Big and fat process when installing some basic pkgs. Dont have facts here but I have seen it too. I had the impression some work went in to get this fixed and improve things. Greetings, stefan -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Community distro
Alan Coopersmith wrote: > > Can Oracle Engineering provide someone like Joerg what he needs to update > > his external SchilliX distro to implement ON snv_145/IPS 134? > > I'm guessing you're talking to me? I have no idea what Joerg needs, but > we've > already provided the community with everything they need to build current ON & > IPS bits, as evidenced by multiple external community members doing so. You > do need a recent build of IPS to build ON builds 136 & later, and X builds > 144 & later, but that's fully available from the pkg-gate sources on the > opensolaris hg repos. Do you have a IPS SVr4 package or do you have a IPS source that compiles on a machine without IPS? Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SchilliX-0.7.0 ready for testing
>Ian asked: >The kernel distribution for Blastwave maybe? That would depend on people like Dennis Clarke. But I did have a little fun in a conceptual build. I'll see if Blastwave (BW) or Genunix will let me post the ON binaries there. Src: -rw-r--r-- 1 kmays csw 83309980 Jul 24 15:33 on-src_b145.tar.bz2 $ uname -a SunOS aerof22 5.11 snv_145 i86pc i386 i86pc $ cat /etc/release SchilliX Community Enterprise Server snv_145 X86 Copyright 2010 Oracle Corporation. All Rights Reserved. Use is subject to license terms. Assembled 24 July 2010 Pre-tested: X.Org X Server 1.7.3 for integration and Nvidia 256.35. Review: Pre-tested: X.Org X Server 1.8.2 for integration and Nvidia 256.35. Main C/C++ compilers to use: Sun Studio 12.1/GCC 4.3.4 (BW) ~ Ken Mays -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Community distro
Ken Mays wrote: > Hi Alan, > > I'm taking a stab at this but I heard Joerg mentioned an issue with moving > beyond ON snv_130. Then, you brought up some things about not being able to > build ON 135+, X b143+, GNOME/JDS, and a few things without IPS being a part > of the distro. So, IPS is definitely on the menu. > > Can Oracle Engineering provide someone like Joerg what he needs to update his > external SchilliX distro to implement ON snv_145/IPS 134? I'm guessing you're talking to me? I have no idea what Joerg needs, but we've already provided the community with everything they need to build current ON & IPS bits, as evidenced by multiple external community members doing so. You do need a recent build of IPS to build ON builds 136 & later, and X builds 144 & later, but that's fully available from the pkg-gate sources on the opensolaris hg repos. -- -Alan Coopersmith-alan.coopersm...@oracle.com Oracle Solaris Platform Engineering: X Window System ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SchilliX-0.7.0 ready for testing
To all those who seemed to have been confused by my reply to Ian's post please note the following: 1) My comments were intended as a response to Ian's assertion of the 'inevitability of IPS'. 2) The example I provided, Nexenta, is pretty strong evidence that incorporation of IPS is not requisite for a successful Open Solars derivative. 3) My comments did not even touch vaguely upon any issues of technical merit of package systems. 4) I was here when Indiana was released. My recollection is that there was quite a bit of uproar about IPS, among other things, specifically in the context of the manner in which Sun played their trump card. Others are certainly free to interpret those events differently. 5) Funny that I'm accused of stirring the pot when the fact of the matter is that Ian used the ruse of a pat on the back as a thinly veiled cover for effectively telling Joerg that he'd better toe the Oracle IPS line. At least that was my read. Hence my response. Your mileage may vary. -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] [distribution-discuss] SchilliX-0.7.0 ready for testing
"C. Bergström" wrote: > >> Which compiler did you use to build onnv-gate? > >> > > > > OS-12.1 > > > That's not a compiler version. I know gcc doesn't boot, but ping me off > list or on irc if you'd like to help test a compiler to replace SS Ot may be calles OSS-12.1, but you should knoe it anyway.. > >> Which version of grub did you include on the ISO and or do you have any > >> other GPLv2 software linking against libc? > >> > > > > > > > To cut to the chase and not armchair lawyer this issue... SFLC has > stated that GPLv2 grub linking to a liberated libc can't be distributed > together.. (eg on an iso) The solution is/was to use grub2 which is > GPLv3 and more clear on this.. Seth Goldberg has done this as well as > one person in the community so it's just a matter of getting their work > integrated. Let me give you the strong advise not to believe in people who claim that the GPLv2 is a non-free license. The claim that GPLv2 software cannot be distributed with other OSS on the same medium is complete nonsense and would make the GPL to be in conflict with the OSS definition: http://www.opensource.org/docs/definition.php that requires a OSS licesense not to restrict other software. Unless you have a written paper to prove your claims, unless this is based on legaly valid quotes on law texts and license texts and unless you publish the paper so it can be reviewed by lawyers, I propose to ignore this claim. BTW: There was a related discussion on OpenSource.org aprox. 8 years ago as the GPL was declared to be a non-OSS license by the OSI until the FSF send a letter that made clear that the GPL has to be interpreted in a way that makes it compliant with the OSS definition. Unfortunately the mail archives from the OSI only go back by 3 years. I recommend you to ask Russ Nelson about the case > There's our patches to libc, packaging and many things.. Please answer > my question and yes all my previous work is public as well as my future > work.. Can you give an URL? Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] [distribution-discuss] SchilliX-0.7.0 ready for testing
2010/7/24 "C. Bergström" : > Joerg Schilling wrote: >> >> "C. Bergström" wrote: >> >> >>> >>> Did you pull the libc work from Garrett and drop that on there? >>> >> >> No, I just created a halfway clean base for starting with the emancipation >> work. >> >> >>> >>> Which compiler did you use to build onnv-gate? >>> >> >> OS-12.1 >> > > That's not a compiler version. I know gcc doesn't boot, but ping me off list > or on irc if you'd like to help test a compiler to replace SS That´s a vague statement. Last time I did this gcc compiled 64bit kernels did boot (sun4u, sun4v, amd64). The 32bit x86 kernel alone was unbootable (and who wants that nowadays, except maybe for embedded use). ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] [distribution-discuss] SchilliX-0.7.0 ready for testing
Joerg Schilling wrote: "C. Bergström" wrote: Did you pull the libc work from Garrett and drop that on there? No, I just created a halfway clean base for starting with the emancipation work. Which compiler did you use to build onnv-gate? OS-12.1 That's not a compiler version. I know gcc doesn't boot, but ping me off list or on irc if you'd like to help test a compiler to replace SS What are you doing for c++ and are the binaries from Solaris/OpenSolaris able to run? I supply the binary libC from Sun. There is absolutely zero documentation on their C++ ABI, but I believe they've implemented the IA64 ABI. It's rumored that with time it'll be the default. Which version of grub did you include on the ISO and or do you have any other GPLv2 software linking against libc? To cut to the chase and not armchair lawyer this issue... SFLC has stated that GPLv2 grub linking to a liberated libc can't be distributed together.. (eg on an iso) The solution is/was to use grub2 which is GPLv3 and more clear on this.. Seth Goldberg has done this as well as one person in the community so it's just a matter of getting their work integrated. Where's the source to your patches? (I thought you have this all on Berlios.. maybe) For b130, I just needed to change two lines of code in order to make ksh93 and another oss work from outside compile with OS-12.1. If people are interested in an updated version of my distro creator, I may work on a general usable set. BTW: do you have results that others may use? There's our patches to libc, packaging and many things.. Please answer my question and yes all my previous work is public as well as my future work.. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] [distribution-discuss] SchilliX-0.7.0 ready for testing
gJoerg Schilling wrote: Hi, today, I put SchilliX-0.7.0 out. ftp://ftp.berlios.de/pub/schillix/ ftp://ftp.berlios.de/pub/schillix/SchilliX-0.7.0.iso.bz2 ftp://ftp.berlios.de/pub/schillix/README.install Changes since SchilliX-0.6.7: - Updated to use OpenSolaris Nevada Build 130 - Updated to use new Schily tools (e.g. cdrtools-3.00) - The history editor in the Bourne shell now supports multi byte character locales. - root now has the initial passwd root - schillix now has the initial passwd schillix Please test an report problems as I am going to use SchilliX-0.7.0 as a base for working on a 100% free and open source based distribution that is able to compile it's own sourcecode. This is usually called "self hosting". Jörg, Did you pull the libc work from Garrett and drop that on there? Which compiler did you use to build onnv-gate? What are you doing for c++ and are the binaries from Solaris/OpenSolaris able to run? Which version of grub did you include on the ISO and or do you have any other GPLv2 software linking against libc? Where's the source to your patches? (I thought you have this all on Berlios.. maybe) Thanks! ./C ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Problem with snv_134 networking in zones
> From what I can see, the problem is resource bound. Either the > CPU can't handle the virtualization, or the NIC driver isn't handling it. And you determined this how? Can you show any dtrace output, or similar? > Whichever it is, the machine is going back to something simpler - > like FreeBSD. I think you might be confusing simpler with familiar. Best of luck. -sean ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SchilliX-0.7.0 ready for testing
Shawn Walker wrote: > On 07/23/10 04:41 PM, Ken Gunderson wrote: > > Why? IPS was shoved down the community's throat in a heavy handed and > > decidedly not FOSS manner. > > Sorry, but that's simply not true. > > The pkg(5) project has been one of the few projects that is actually > very open. It was the first to use defect.opensolaris.org for > bugtracking, it's licensed under the CDDL, and at the moment any > contributor (even external ones) can get commit access upon approval by > the project team members. The basic OSS rules are: - release early - release often - port to many platforms. But there was no visibility and "testability" on SXCE and for this reason, it missed most users I know. For now, IPS is used by one OpenSolaris distro only and this distro published the last binary release nearly one year ago. Writing and publishing successful OSS is not only producing code. You need to convince the users also with facts. The best way this can be done is by making the software available everywhere (and creators of distributions have a big change with getting new users by including software), but IPS did not appear on SXCE... Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SchilliX-0.7.0 ready for testing
On Sat, Jul 24, 2010 at 2:11 PM, Robert Milkowski wrote: > On 24/07/2010 02:39, Shawn Walker wrote: >> >> On 07/23/10 04:41 PM, Ken Gunderson wrote: >>> >>> Why? IPS was shoved down the community's throat in a heavy handed and >>> decidedly not FOSS manner. >> >> Sorry, but that's simply not true. >> >> The pkg(5) project has been one of the few projects that is actually very >> open. It was the first to use defect.opensolaris.org for bugtracking, it's >> licensed under the CDDL, and at the moment any contributor (even external >> ones) can get commit access upon approval by the project team members. >> >> It has active, external contributors to the project (including myself at >> one point a few years ago before I was employed for the project), and is one >> of the few to push almost all design and development discussions onto a >> public os.org mailing list. >> >> Remember that this community and the projects that provide the basis for >> various OpenSolaris distributions remain largely a meritocracy -- those that >> do the work get to make the decisions. >> >> There are plenty of open source projects that have made decisions >> unpopular with their user community. That doesn't make those projects any >> less FOSS, nor does it justify claims of "forcing" something on a community. >> > I have yet to see any large FOSS project which does some kind of community > voting, or something like that. > I agree with you 100%. > > The only problem I can see in regards to how IPS was introduced that perhaps > more explanation to the community at the project inception about why IPS and > not something else would do good... a little bit more open dialog with the > community before you start coding on why IPS, what are its design goals, > etc. would help here. As most of it happened well after you made your mind > and started coding therefor some people felt being disregarded. But after > this you are right - it's probably one of the most open projects here. > > But honestly, it is only a relatively small group of people here who are > complaining about IPS design - but then they do not propose (code please) > any real alternative. Most of the end-users I know do like IPS actually - > they used to complain that it is slow (which is still the case sometimes) or > that it consumes too much memory which is especially visible on notebooks, > etc. But generally they do like it and one can see how it has been improving > over time. > To stir the pot here, since we are discussing a "Community Distro" as opposed to a SUN/Oracle distro, IPS when used remotely from halfway across the world has large performance issues. For example in Bangalore I personally know no one outside the SUN India office who have successfully updated packages on their OpenSolaris installations let alone do an image-update in a sane amount of time. In comparison things like YUM or Apt-Get are reasonable even over 512Kbps. Having said that I am not going to make vague statements. I will do my own testing over a 2Mbps broadband link and post measurements. Regards, Moinak. -- http://www.belenix.org/ http://moinakg.wordpress.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Problem with snv_134 networking in zones
I can't see how virtualising a NIC can have a lesser resource impact on an already challenged machine than using aliases. >From what I can see, the problem is resource bound. Either the CPU can't >handle the virtualization, or the NIC driver isn't handling it. Whichever it is, the machine is going back to something simpler - like FreeBSD. Quite unfortunate though, since I did have plans for it running OSol. Oh well... Next year, maybe :) -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Community distro
On Sat, Jul 24, 2010 at 4:01 PM, Joerg Schilling wrote: > Moinak Ghosh wrote: > >>I am taking guesses here but I suspect this boot performance may >>be relative. AFAIK SMF does not restrict the extent of concurrency. >>So if there are 100 non-interdependent services ready to be started >>at the current graph state then it will start all 100 concurrently. This >>might be good for multi-way big-iron boxes but can be at the >>diminishing returns point for smaller workstations and laptops where >>there is neither the disk/memory bandwidth and nor the core-count to >>handle such a load. In this case a feature to have a graceful reduction >>in concurrency level in SMF should help. > > It may even be seek times are the main problem... > True. Too much concurrency on spinning platters can cause seek time bottlenecks. Possibly an iostat log during boot can provide some clues. If true then the situation will be better with SSDs. Regards, Moinak. -- http://www.belenix.org/ http://moinakg.wordpress.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Community distro
Hi Alan, I'm taking a stab at this but I heard Joerg mentioned an issue with moving beyond ON snv_130. Then, you brought up some things about not being able to build ON 135+, X b143+, GNOME/JDS, and a few things without IPS being a part of the distro. So, IPS is definitely on the menu. Can Oracle Engineering provide someone like Joerg what he needs to update his external SchilliX distro to implement ON snv_145/IPS 134? ~ Ken Mays P.S. 'To infinity and beyond...' - Toy Story -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Community distro
Moinak Ghosh wrote: >I am taking guesses here but I suspect this boot performance may >be relative. AFAIK SMF does not restrict the extent of concurrency. >So if there are 100 non-interdependent services ready to be started >at the current graph state then it will start all 100 concurrently. This >might be good for multi-way big-iron boxes but can be at the >diminishing returns point for smaller workstations and laptops where >there is neither the disk/memory bandwidth and nor the core-count to >handle such a load. In this case a feature to have a graceful reduction >in concurrency level in SMF should help. It may even be seek times are the main problem... Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SchilliX-0.7.0 ready for testing
On 24/07/2010 02:39, Shawn Walker wrote: On 07/23/10 04:41 PM, Ken Gunderson wrote: Why? IPS was shoved down the community's throat in a heavy handed and decidedly not FOSS manner. Sorry, but that's simply not true. The pkg(5) project has been one of the few projects that is actually very open. It was the first to use defect.opensolaris.org for bugtracking, it's licensed under the CDDL, and at the moment any contributor (even external ones) can get commit access upon approval by the project team members. It has active, external contributors to the project (including myself at one point a few years ago before I was employed for the project), and is one of the few to push almost all design and development discussions onto a public os.org mailing list. Remember that this community and the projects that provide the basis for various OpenSolaris distributions remain largely a meritocracy -- those that do the work get to make the decisions. There are plenty of open source projects that have made decisions unpopular with their user community. That doesn't make those projects any less FOSS, nor does it justify claims of "forcing" something on a community. I have yet to see any large FOSS project which does some kind of community voting, or something like that. I agree with you 100%. The only problem I can see in regards to how IPS was introduced that perhaps more explanation to the community at the project inception about why IPS and not something else would do good... a little bit more open dialog with the community before you start coding on why IPS, what are its design goals, etc. would help here. As most of it happened well after you made your mind and started coding therefor some people felt being disregarded. But after this you are right - it's probably one of the most open projects here. But honestly, it is only a relatively small group of people here who are complaining about IPS design - but then they do not propose (code please) any real alternative. Most of the end-users I know do like IPS actually - they used to complain that it is slow (which is still the case sometimes) or that it consumes too much memory which is especially visible on notebooks, etc. But generally they do like it and one can see how it has been improving over time. -- Robert Milkowski http://milek.blogspot.com ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SchilliX-0.7.0 ready for testing
On 24/07/2010 00:17, Ian Collins wrote: I really do think you should accept the inevitable and accept IPS, otherwise SchilliX runs the risk of becoming an evolutionary dead end. If any OpenSolaris distributions are to survive post-Oracle, they must be able to share packages with each other and ultimately with Oracle's Solaris. A disparate packaging scheme shows a fractured community. We should at least strive to appear united! 100% agree -- Robert Milkowski http://milek.blogspot.com ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org