[osol-discuss] Solaris-windows, samba, strange behaviour
Hello! I am using the SUNWsmba-package. 1. Transfers from Solaris-winxp/win7 works great. (some occasional slowdowns on the xp, but the machine is 7 years old so...) solaris-xp http://basse.host22.com/junk/solaris_to_xp.jpg solaris-win7 http://basse.host22.com/junk/greenshot_2009-12-14_11-04-01.png 2. transfers between my winxp-win7 machines works great 3. But transfers from win7/winxp-solaris gets a strange behaviour where the transfer stalls every now and then. win xp-solaris: http://basse.host22.com/junk/xp_to_solaris.JPG win7 - solaris: http://basse.host22.com/junk/greenshot_2009-12-14_10-34-52.png I also tried to boot linux, debian, on one machine, and debian-win7/xp works as usual, so it´s definently something at the Solaris machine... Another strange thing to notice is that the Win7-solaris speeds seems twice as big as the solaris-win7 speeds... After 1 week of this, and trying to solve it, I´m out of ideas. Help would be greatly appreciated. /whopp -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Some Why?-Questions
You're getting to see the process from the slaughterhouse through the kitchen, instead of just getting the steak delivered on a plate when it's fully cooked like you did before - it's going to be messy, but hopefully we'll end up with a better product in the end. And that's perfectly fine, great even, no problem with that at all! What makes me personally extremely angry and frustrated is the level of *aggresive* marketing and promotion of OpenSolaris as the be-all, end-all, the next big thing, THE Solaris.Next and use it today!, and then when one does actually attempt to use it and finds all these deficiencies, only THEN do the excuses start: [...] Well, I've watched this thread go by... we've been through this, on this list and others, a number of times. I understand your frustration, as I have been feeling it myself. Yes, Sun has made two big mistakes: Implementing IPS in Python, and ditching scripting capability in the packages. I'm sure these seemed like good engineering decisions way back at the drawing board. But they're a disaster in terms of marketing, installed base, and usability. But Sun has always been engineering rather than marketing driven, and frankly, that's why I like them. :-) In the end, we need to move on in a constructive way. They're not going to axe IPS and replace it with SysV packages or anything else. It's here to stay. Same goes for AI and OpenSolaris as a whole. What we need now is best practices guides/cookbooks/howtos/whatever on how to overcome these deficiencies: How to design packages so that they don't hamper IPS performance, how to create SMF services that do the jobs previously done by preinstall/postinstall and CAS. I want blueprints, white papers, code samples, etc. etc. Look at the Apple knowledgebase to see it done right. While ranting and raving feels good immediately, working on a good documentation base that includes real life working samples of well-defined packages will be more worthwhile in the long run. Because I do want that database that springs to life just out of a pkgadd... errr... pkg install. :-) Regards -- Volker -- Volker A. Brandt Consulting and Support for Sun Solaris Brandt Brandt Computer GmbH WWW: http://www.bb-c.de/ Am Wiesenpfad 6, 53340 Meckenheim Email: v...@bb-c.de Handelsregister: Amtsgericht Bonn, HRB 10513 Schuhgröße: 45 Geschäftsführer: Rainer J. H. Brandt und Volker A. Brandt ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Jumbo frames on rge
Which build are you seeing this with/from? .s -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Jumbo frames on rge
Which build are you seeing this with/from? [cheeky] ~ # cat /etc/release Solaris Express Community Edition snv_125 X86 Copyright 2009 Sun Microsystems, Inc. All Rights Reserved. Use is subject to license terms. Assembled 05 October 2009 [cheeky] ~ # uname -a SunOS cheeky 5.11 snv_128 i86pc i386 i86pc So it is SX:CE snv_125, bfu'd to 128. .s Chavdar Ivanov -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] snv_111b x86 crashing
melbogia wrote: There is no panic or syncing filesystems in /var/adm/messages. Here is what I see after I execute the command mkfile 100g /datapool/testfile I was on the machine at 1600 hours yesterday and the next entry after that at 8:43 AM is system startup information, it seems it can't even write anything to /var/adm/messages before it reboots? Also this may be relevant, the datapool pool doesn't exist after it comes up from crash. I have to do a zpool import -f datapool to import it. This may be multiple issues. The fact a reboot happens with no trace in /var/adm/messages is somewhat worrying to me. However, we need to frame the issue. Things like: When you were using the system on the 10th, how were you connected to it? Were you logged in remotely (via ssh/telnet/etc...)? Where you on the console? (graphics console, or text mode)? What did you see on the screen around 16:19 ? Can you check any alternate boot environments, and make sure you are running from the BE you expect to be running from? e.g. did you have a pending beadm activate, and the reboot simply booted into the new environment? Also, are there any known hardware reset issues for your platform? There have been bugs occasionally in the past where a driver will cause a system-wide reset without going through the panic mechanism. I can't recall any, but do remember one or two cropping up occasionally. Dec 11 08:43:36 dirt zfs: [ID 427000 kern.warning] WARNING: pool 'datapool' could not be loaded as it was last accessed by another system (host: dirt hostid: 0x409a4c). See: http://www.sun.com/msg/ZFS-8000-EY Whilst I can't comment on this line myself, it does (IMO) explain why you had to force import the pool after the reboot. Assuming there is only one host with name dirt, then this would seem to be a bug. Regards, Brian -- Brian Ruthven Solaris Revenue Product Engineering Sun Microsystems UK Sparc House, Guillemont Park, Camberley, GU17 9QG ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Some Why?-Questions
Volker A. Brandt wrote: Yes, Sun has made two big mistakes: Implementing IPS in Python, and ditching scripting capability in the packages. I'm sure these seemed I continue to see assertions that pkg(5) should not have been written in python with little justification for this claim. Significant portion of softwares in use at large software installations for high-traffic providers such as Google are written in Python. It is important to keep in mind that algorithms generally matter more than choice of programming language. Cheers, -- Shawn Walker ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] snv_111b x86 crashing
melbogia wrote: There is no panic or syncing filesystems in /var/adm/messages. Here is what I see after I execute the command mkfile 100g /datapool/testfile I was on the machine at 1600 hours yesterday and the next entry after that at 8:43 AM is system startup information, it seems it can't even write anything to /var/adm/messages before it reboots? Also this may be relevant, the datapool pool doesn't exist after it comes up from crash. I have to do a zpool import -f datapool to import it. This may be multiple issues. The fact a reboot happens with no trace in /var/adm/messages is somewhat worrying to me. However, we need to frame the issue. Things like: When you were using the system on the 10th, how were you connected to it? Were you logged in remotely (via ssh/telnet/etc...)? Where you on the console? (graphics console, or text mode)? What did you see on the screen around 16:19 ? I was connected to the server via ssh and I didn't see anything at 16:19. an you check any alternate boot environments, and make sure you are running from the BE you expect to be running from? e.g. did you have a pending beadm activate, and the reboot simply booted into the new environment? There is only one BE in the system r...@dirt:~# beadm list BE Active Mountpoint Space Policy Created -- -- -- - -- --- opensolaris NR / 137.25G static 2009-12-09 22:05 Also, are there any known hardware reset issues for your platform? There have been bugs occasionally in the past where a driver will cause a system-wide reset without going through the panic mechanism. I can't recall any, but do remember one or two cropping up occasionally. I am not sure about the hardware reset issues, I'll poke around and see what I find. But there is another machine with the Opensolaris 2009.06 as well, it has the exact same hardware with the exception of hard disks. It has been working fine. Dec 11 08:43:36 dirt zfs: [ID 427000 kern.warning] WARNING: pool 'datapool' could not be loaded as it was last accessed by another system (host: dirt hostid: 0x409a4c). See: http://www.sun.com/msg/ZFS-8000-EY Whilst I can't comment on this line myself, it does (IMO) explain why you had to force import the pool after the reboot. Assuming there is only one host with name dirt, then this would seem to be a bug. that is correct, there is only one host with that name in the environment. -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] snv_111b x86 crashing
melbogia wrote: melbogia wrote: There is no panic or syncing filesystems in /var/adm/messages. Here is what I see after I execute the command mkfile 100g /datapool/testfile I was on the machine at 1600 hours yesterday and the next entry after that at 8:43 AM is system startup information, it seems it can't even write anything to /var/adm/messages before it reboots? Also this may be relevant, the datapool pool doesn't exist after it comes up from crash. I have to do a zpool import -f datapool to import it. This may be multiple issues. The fact a reboot happens with no trace in /var/adm/messages is somewhat worrying to me. However, we need to frame the issue. Things like: When you were using the system on the 10th, how were you connected to it? Were you logged in remotely (via ssh/telnet/etc...)? Where you on the console? (graphics console, or text mode)? What did you see on the screen around 16:19 ? I was connected to the server via ssh and I didn't see anything at 16:19. I can't imagine you saw absolutely nothing from your ssh session. I'm guessing that it hung, and eventually gave some error like timeout or connection reset by peer or something? Even if you reset it, did you try ssh again, ping, etc...? What were the results here? Did you go (if able) to the system itself? What state was it in? etc... It wasn't a power cut was it? ;-) I am not sure about the hardware reset issues, I'll poke around and see what I find. But there is another machine with the Opensolaris 2009.06 as well, it has the exact same hardware with the exception of hard disks. It has been working fine. That's kind of my point. If there is an identical system (and beware - they are rarely exactly identical with no differences whatsoever), and it is functioning fine, then (assuming similar usage patterns) it may help point to a hardware/firmware issue, etc... Don't get me wrong - I'm not saying your hardware is bust - just something to bear in mind if the fault always stays with the one system and no other system is affected. Dec 11 08:43:36 dirt zfs: [ID 427000 kern.warning] WARNING: pool 'datapool' could not be loaded as it was last accessed by another system (host: dirt hostid: 0x409a4c). See: http://www.sun.com/msg/ZFS-8000-EY Whilst I can't comment on this line myself, it does (IMO) explain why you had to force import the pool after the reboot. Assuming there is only one host with name dirt, then this would seem to be a bug. that is correct, there is only one host with that name in the environment. You might want to ask this question on zfs-discuss, and give them the background. Just a thought (assuming the system will let you do this), who else was logged in at the time? Could somebody have done a forced export of all zpools on the system? Could this explain why the messages file stopped being written to? Regards, Brian -- Brian Ruthven Solaris Revenue Product Engineering Sun Microsystems UK Sparc House, Guillemont Park, Camberley, GU17 9QG ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Some Why?-Questions
Shawn Walker writes: Volker A. Brandt wrote: Yes, Sun has made two big mistakes: Implementing IPS in Python, and ditching scripting capability in the packages. I'm sure these seemed I continue to see assertions that pkg(5) should not have been written in python with little justification for this claim. Hmmm... I will readily admit that you're working on fixing the performance issues, and improving overall efficiency, so the situation is better now than it was when pkg(5) was first released into the wild. However, there is also the fact that Sun had already committed to a scripting language, Perl. There was a statement that Perl was a core part of Solaris and would always be present on the miniroot. (I am not saying that Perl would be markedly faster here.) Significant portion of softwares in use at large software installations for high-traffic providers such as Google are written in Python. True. So? Many big things are written in JavaScript, PHP, etc. I for one certainly wouldn't want to see pkg(5) in PHP. :-) It is important to keep in mind that algorithms generally matter more than choice of programming language. Quite right. But this is all water under the bridge. I would like to see a sample package that delivers an SMF service that does some arbitrary post-install task (like configuring and starting Sybase :-) and then removes the SMF service from the system. That's what people need, not theoretical discussions over Python ./. Perl ./. C. Yes, it's on my ToDo list... it's not *that* hard. The only problem is that lacking a Sun-endorsed sample, people will reinvent hundreds of differently-shaped wheels. Regards -- Volker -- Volker A. Brandt Consulting and Support for Sun Solaris Brandt Brandt Computer GmbH WWW: http://www.bb-c.de/ Am Wiesenpfad 6, 53340 Meckenheim Email: v...@bb-c.de Handelsregister: Amtsgericht Bonn, HRB 10513 Schuhgröße: 45 Geschäftsführer: Rainer J. H. Brandt und Volker A. Brandt ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Some Why?-Questions
Volker A. Brandt wrote: Shawn Walker writes: Volker A. Brandt wrote: Yes, Sun has made two big mistakes: Implementing IPS in Python, and ditching scripting capability in the packages. I'm sure these seemed I continue to see assertions that pkg(5) should not have been written in python with little justification for this claim. Hmmm... I will readily admit that you're working on fixing the performance issues, and improving overall efficiency, so the situation is better now than it was when pkg(5) was first released into the wild. However, there is also the fact that Sun had already committed to a scripting language, Perl. There was a statement that Perl was a core part of Solaris and would always be present on the miniroot. (I am not saying that Perl would be markedly faster here.) And there can be only one? Doesn't that mean perl was also a mistake since Sun had already committed to sh as a scripting language available on the miniroot? (Just taking your argument to its logical conclusion - I'm a heavy perl user myself, so think both perl python have their place in the core OS.) -- -Alan Coopersmith- alan.coopersm...@sun.com Sun Microsystems, Inc. - X Window System Engineering ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] ZFS Questions
I've been working trying to create a minimal opensolaris VirtualBox appliance and it looks like by default some sort of mirroring is enabled within the filesystem. Is it possible/easy to disable this. The appliance is just supposed to be a static development environment for the users and will be backed up separately anyway. -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Some Why?-Questions
I really do not see need for Python, Perl and other scripting languagesto be placed in SunOS. I believe that SUN has to buy Design Patternsand Refactoring to patterns books along with The art of computer scienceand share to their developers to continue developing perfect such importantpiece of software. Also, I would like to say that I believe that SPARC will stay on its lineto be almost the best microprocessor available on the market. If we alltransit to X86 all our knowledge in computer science will worth nothing. Uros --- Every kind of peaceful cooperation among men is primarily based on mutual trust and only secondarily on institutions such as courts of justice and police. - Albert Einstein (1879 - 1955) Date: Mon, 14 Dec 2009 09:47:22 -0800 From: alan.coopersm...@sun.com To: v...@bb-c.de CC: opensolaris-discuss@opensolaris.org Subject: Re: [osol-discuss] Some Why?-Questions Volker A. Brandt wrote: Shawn Walker writes: Volker A. Brandt wrote: Yes, Sun has made two big mistakes: Implementing IPS in Python, and ditching scripting capability in the packages. I'm sure these seemed I continue to see assertions that pkg(5) should not have been written in python with little justification for this claim. Hmmm... I will readily admit that you're working on fixing the performance issues, and improving overall efficiency, so the situation is better now than it was when pkg(5) was first released into the wild. However, there is also the fact that Sun had already committed to a scripting language, Perl. There was a statement that Perl was a core part of Solaris and would always be present on the miniroot. (I am not saying that Perl would be markedly faster here.) And there can be only one? Doesn't that mean perl was also a mistake since Sun had already committed to sh as a scripting language available on the miniroot? (Just taking your argument to its logical conclusion - I'm a heavy perl user myself, so think both perl python have their place in the core OS.) -- -Alan Coopersmith- alan.coopersm...@sun.com Sun Microsystems, Inc. - X Window System Engineering ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org _ Windows Live Hotmail: Your friends can get your Facebook updates, right from Hotmail®. http://www.microsoft.com/middleeast/windows/windowslive/see-it-in-action/social-network-basics.aspx?ocid=PID23461::T:WLMTAGL:ON:WL:en-xm:SI_SB_4:092009___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Some Why?-Questions
UNIX admin wrote: Because if we want to be ready for the future, we must now maintain two sets of packages for every component - one for the enterprise, which is what feeds us and pays the bills, one for being ready for the future. And if it wasn't IPS, then it would be some other feature in Solaris 11 that would make you have to choose between the least common denominator and supporting all the new features. This is the same dilemma every OS provides with new releases - Do I have one common binary for all versions, or customized ones that take advantage of newer features in newer releases? - it could be Solaris 8 vs. 10 (see the blastwave dilemma on linking to open source libraries in the OS there for instance) or Windows XP vs. 7. But that costs tremendous amounts of effort and money; it's very expensive. pkgadd(1M) could have been incrementally improved with the backgraph algorithm in AWK and the C programming language books which the make(1) tool also uses, why wasn't this done instead? pkgadd(1M) could have been incrementally improved, based on pkgtrans(1), to have knowledge of true package clusters instead of the loose package metacluster (like SUNWCall), why wasn't this done? pkgadd(1M)'s capability to install packages via http:// protocol could have been extended further, coupled with the dependency resolution algorithm, to automatically install any and all needed packages over the network, like yum install and pkg_get(1M) do; why wasn't this done? Why did Sun have to create ZFS instead of just extending UFS more? Why did Sun have to create SMF instead of just extending init.d scripts more? Why did Sun have to move to GNOME instead of just extending CDE more? Why did Sun have to move to SVR4 instead of just extending SunOS 4 more? Why did Sun have to create SPARC instead of just building more 680x0 machines? Change is inevitable in IT - sometimes the amount of change you need to make is so great that replacement is the most viable option (allowing side-by-side implementations during the transition and for customers to transition at their own speed).Some of these have been more painful than others, but the end result was better than hacking new features into an old design they didn't fit into. I understand you might not have the answers to these questions; but surely someone inside of Sun Microsystems knows! Yes, and Stephen and Bart have explained it quite a bit - if everything they've said and written about their investigations of the options and the requirements they gathered from the various major enterprise customers they talked to hasn't convinced you, my third-hand repeating of what they said isn't going to help. -- -Alan Coopersmith- alan.coopersm...@sun.com Sun Microsystems, Inc. - X Window System Engineering ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Some Why?-Questions
Alan Coopersmith writes: Volker A. Brandt wrote: However, there is also the fact that Sun had already committed to a scripting language, Perl. There was a statement that Perl was a core part of Solaris and would always be present on the miniroot. (I am not saying that Perl would be markedly faster here.) And there can be only one? Doesn't that mean perl was also a mistake since Sun had already committed to sh as a scripting language available on the miniroot? (Just taking your argument to its logical conclusion Good point. I actually would have preferred a C implementation for pkg(5). Anyway, water under the bridge. - I'm a heavy perl user myself, so think both perl python have their place in the core OS.) I certainly wouldn't advocate kicking out Python. :-) What gets lost in this discussion is the need for a bridge over the gap between you Sun engineers in your ivory tower designing pure and well-defined systems and us consultants and software developers needing to implement automation during system installation in some reasonable reproducible way. Regards -- Volker -- Volker A. Brandt Consulting and Support for Sun Solaris Brandt Brandt Computer GmbH WWW: http://www.bb-c.de/ Am Wiesenpfad 6, 53340 Meckenheim Email: v...@bb-c.de Handelsregister: Amtsgericht Bonn, HRB 10513 Schuhgröße: 45 Geschäftsführer: Rainer J. H. Brandt und Volker A. Brandt ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Some Why?-Questions
Hi Uros! I really do not see need for Python, Perl and other scripting languagesto be placed in SunOS. I believe that SUN has to buy Design Patternsand Refactoring to patterns books along with The art of computer scienceand share to their developers to continue developing perfect such importantpiece of software. I think you are exaggerating a little. :-) Also, I would like to say that I believe that SPARC will stay on its lineto be almost the best microprocessor available on the market. If we alltransit to X86 all our knowledge in computer science will worth nothing. Now you are exaggerating more than a little. :-))) Let's look forward and make OpenSolaris and IPS and SMF a nice system on all platforms. Regards -- Volker -- Volker A. Brandt Consulting and Support for Sun Solaris Brandt Brandt Computer GmbH WWW: http://www.bb-c.de/ Am Wiesenpfad 6, 53340 Meckenheim Email: v...@bb-c.de Handelsregister: Amtsgericht Bonn, HRB 10513 Schuhgröße: 45 Geschäftsführer: Rainer J. H. Brandt und Volker A. Brandt ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Some Why?-Questions
Hi Volker, I really hope that I exaggerate but when I spoke with one ofSenior Engineers of Sandy Bridge architecture duringIDF this September I realized where we go. To be worse,he also agreed, too, even he is working for Intel. The bestin Intel is their Physical Electronics Research Department,and moving to smaller electronic components each two years. Architecture is almost broken - they become hostages oftheir own success - X86 architecture. Same is happening with Microsoft - they also become hostagesof their own success - Windows NT. They are trying to deploy the latest innovations inside chip/kernel,to improve their products but not to change interface. Parts ofWindows 7 has been completely rewritten to be faster just notto be ABI compatible with older versions of Windows productsline. At some point in future this has to be stopped. This is whatI'm talking about. Software engineers (including me, but tryingto improve) live on account of improvement of computer architecturewhich in turn live on account of innovations in physical electronics,which does its job the best of all three industries. Now I'll go a little bit into economy. Why people by X86?Because they are cheap! Why people buy Windows Serverplatform? Because it was cheap! So this is one of examplesof inefficiency or deviation of market economy, of which I'mby the way supporter but not fundamentalist. If anyone care about that I could give list of books of well-knownResearchers, Professors, etc., to read about that. It wasalready written and well studied but industry/economy usuallyhas to come unfortunately to the dead to see mistakes. Then,price for recovering is usually very high. Hope I clarified my perspective, at least a little bit. Regards - Nedic --- Every kind of peaceful cooperation among men is primarily based on mutual trust and only secondarily on institutions such as courts of justice and police. - Albert Einstein (1879 - 1955) From: v...@bb-c.de Date: Mon, 14 Dec 2009 19:10:59 +0100 To: ur...@live.com CC: opensolaris-discuss@opensolaris.org Subject: Re: [osol-discuss] Some Why?-Questions Hi Uros! I really do not see need for Python, Perl and other scripting languagesto be placed in SunOS. I believe that SUN has to buy Design Patternsand Refactoring to patterns books along with The art of computer scienceand share to their developers to continue developing perfect such importantpiece of software. I think you are exaggerating a little. :-) Also, I would like to say that I believe that SPARC will stay on its lineto be almost the best microprocessor available on the market. If we alltransit to X86 all our knowledge in computer science will worth nothing. Now you are exaggerating more than a little. :-))) Let's look forward and make OpenSolaris and IPS and SMF a nice system on all platforms. Regards -- Volker -- Volker A. Brandt Consulting and Support for Sun Solaris Brandt Brandt Computer GmbH WWW: http://www.bb-c.de/ Am Wiesenpfad 6, 53340 Meckenheim Email: v...@bb-c.de Handelsregister: Amtsgericht Bonn, HRB 10513 Schuhgröße: 45 Geschäftsführer: Rainer J. H. Brandt und Volker A. Brandt _ Windows Live Hotmail: Your friends can get your Facebook updates, right from Hotmail®. http://www.microsoft.com/middleeast/windows/windowslive/see-it-in-action/social-network-basics.aspx?ocid=PID23461::T:WLMTAGL:ON:WL:en-xm:SI_SB_4:092009___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] OpenSolaris hard disk hot plug?
Hi, I have a Sun Fire X2270, I have enabled hot plug option in BIOS configuration. I have running the server with OpenSolaris 2009.06, when I insert into the SATA slot 1 a new hard disk I expect that a char device to appear in /dev/rdsk/ but nothing happens. How can I enable hot plug in OpenSolaris or how this works? Thank you. xavier. -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] OSOL SNV_130: You've come a long way, baby...
Although OSOL 2010.03 is still a few months away, we are starting to get down to some of the main features we may see in the final product release. Interesting tidbits: 1. The new IPS repositories and packaging requirements. 2. Sun has moved to providing the Nvidia video driver and graphic cards which are some of the premier graphic visualization products on the market. 3. The move to ZFS and providing ZFS boot capability. Also, Time Slider. 4. SongBird music client 5. The Boomer OSS4 audio framework 6. Xorg 7.5 7. Adobe Reader 9 Sun is trying to push a few envelopes in providing a hip and current operating systems on both the server and desktop client. We are still a week away from OSOL SNV_130. So time to jot down the notes on what we'd like to see and improvements Sun needs to make before they 'nail the coffin' shut. ~ Ken Mays ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] OpenSolaris hard disk hot plug?
it's not how it works, read the related manual in docs.sun.com http://docs.sun.com/app/docs/doc/817-3814/6mjcp0qp9?a=view here is a hint, you'll need cfgadm On Mon, Dec 14, 2009 at 4:24 PM, Xavier Callejas xav...@wflogistics.com wrote: Hi, I have a Sun Fire X2270, I have enabled hot plug option in BIOS configuration. I have running the server with OpenSolaris 2009.06, when I insert into the SATA slot 1 a new hard disk I expect that a char device to appear in /dev/rdsk/ but nothing happens. How can I enable hot plug in OpenSolaris or how this works? Thank you. xavier. -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] tcpdump -A equivalent in snoop?
Chavdar Ivanov wrote: $ pfexec pkg set-publisher -O http://pkg.opensolaris.org/contrib/ contrib $ pfexec pkg install tcpdump will show up in the dev repo in build 129. What did I miss? Updated from 126 to 129 build, installed SUNWtcpdump. Now tcpdump complains that bpf is missing: $ pfexec tcpdump -i e1000g0 tcpdump: (cannot open device) /dev/bpf: No such file or directory It is really missing, but why? -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] tcpdump -A equivalent in snoop?
On Mon, 2009-12-14 at 12:46 -0800, Alexander wrote: Chavdar Ivanov wrote: $ pfexec pkg set-publisher -O http://pkg.opensolaris.org/contrib/ contrib $ pfexec pkg install tcpdump will show up in the dev repo in build 129. What did I miss? Updated from 126 to 129 build, installed SUNWtcpdump. Now tcpdump complains that bpf is missing: $ pfexec tcpdump -i e1000g0 tcpdump: (cannot open device) /dev/bpf: No such file or directory It is really missing, but why? Because libpcap now uses BPF instead of libdlpi, and the libpcap package is missing a dependency on the BPF package. This is the following bug: http://defect.opensolaris.org/bz/show_bug.cgi?id=13279 The workaround is to install SUNWpacket. -Seb ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Some Why?-Questions
Volker A. Brandt wrote: Alan Coopersmith writes: Volker A. Brandt wrote: However, there is also the fact that Sun had already committed to a scripting language, Perl. There was a statement that Perl was a core part of Solaris and would always be present on the miniroot. (I am not saying that Perl would be markedly faster here.) And there can be only one? Doesn't that mean perl was also a mistake since Sun had already committed to sh as a scripting language available on the miniroot? (Just taking your argument to its logical conclusion Good point. I actually would have preferred a C implementation for pkg(5). % find . -name '*.c' ./util/misc/extract_hostid.c ./util/distro-import/ksh-wrapper.c ./brand/support.c ./modules/actions/_actions.c ./modules/arch.c ./modules/pspawn.c ./modules/liblist.c ./modules/elf.c ./modules/elfextract.c ./modules/solver/py_solver.c ./modules/solver/solver.c The parts that benefit from being in C are in C. What gets lost in this discussion is the need for a bridge over the gap between you Sun engineers in your ivory tower designing pure and well-defined systems and us consultants and software developers needing to implement automation during system installation in some reasonable reproducible way. Isn't that why we have opensolaris and the community discussions? -- -Alan Coopersmith- alan.coopersm...@sun.com Sun Microsystems, Inc. - X Window System Engineering ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] OSOL SNV_130: You've come a long way, baby...
ken mays wrote: Although OSOL 2010.03 is still a few months away, we are starting to get down to some of the main features we may see in the final product release. We are still a week away from OSOL SNV_130. So time to jot down the notes on what we'd like to see and improvements Sun needs to make before they 'nail the coffin' shut. For most consolidations, code freeze for snv_130 was a week ago, and the final packages (after respins for any issues found in QA over the last week) were due a few hours ago. We're mostly in feature freeze now for 2010.03, so there's room for RFE's and smaller things, but no more major upgrades or projects until after 2010.03 for most of the OS. (Things like IPS outside the main consolidations operate on different schedules, and have a later freeze.) -- -Alan Coopersmith- alan.coopersm...@sun.com Sun Microsystems, Inc. - X Window System Engineering ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] tcpdump -A equivalent in snoop?
Thanks... tcpdump is much more habitual for me... What did I miss? Updated from 126 to 129 build, installed SUNWtcpdump. Now tcpdump complains that bpf is missing: $ pfexec tcpdump -i e1000g0 tcpdump: (cannot open device) /dev/bpf: No such file or directory It is really missing, but why? Because libpcap now uses BPF instead of libdlpi, and the libpcap package is missing a dependency on the BPF package. This is the following bug: http://defect.opensolaris.org/bz/show_bug.cgi?id=13279 The workaround is to install SUNWpacket. -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Some Why?-Questions
Volker A. Brandt wrote: ... But this is all water under the bridge. I would like to see a sample package that delivers an SMF service that does some arbitrary post-install task (like configuring and starting Sybase :-) and then removes the SMF service from the system. That's what people need, not theoretical discussions over Python ./. Perl ./. C. Yes, it's on my ToDo list... it's not *that* hard. The only problem is that lacking a Sun-endorsed sample, people will reinvent hundreds of differently-shaped wheels. As we get closer towards Solaris Next, we recognize that samples, how-to's, and so on are requirements and fully expect to provide a lot of transition aids such as that. Contributions from users such as yourself would be very helpful. Dave ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] OpenSolaris hard disk hot plug?
Ignacio Marambio Catán wrote: here is a hint, you'll need cfgadm Thank you very much! that it's. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] snv_111b x86 crashing
When my ssh connection got stuck I didn't wait for it give me a connection reset by peer error and promptly proceeded to disconnect my connection using the ssh disconnect key pattern, i.e ENTER ~ . I then connected over from serial console and saw machine BIOS messages as they show up when the machine is being restarted. Our serial telnet servers keep a log of the last messages on serial and I just check those. I installed opensolaris over serial using Automated Installer and these are the last messages, which is basically showing OS being installed: 066 Download: SUNWgnome-cd ... Done 066 066 Download: SUNWlpr-cmds ... Done 066 066 Download: SUNWgnu-coreutils ... Done 066 066 Download: SUNWipkg-um ... Done 066 066 Download: SUNWgetting-started-l10n-extra ... Done 066 066 Download: SUNWislcc ... Done 066 066 Download: SUNWtoo ... Done 066 066 Download: SUNWroute ... Done 066 066 Download: SUNWxcursor-themes ... Done 066 066 Download: SUNWpcan ... Done 066 066 Download: SUNWgsed ... Done 066 066 Download: SUNWgnome-l10nmessages-ko ... Done 066 066 Download: SUNWthunderbirdl10n-ru-RU ... Done 066 066 Download: SUNWgetting-started-l10n-zhHK ... Done 066 066 Download: SUNWopensolaris-welcome ... Done 066 066 Download: SUNWssh ... Done 066 066 Download: SUNWthunderbirdl10n-zh-TW ... Done 066 066 Download: SUNWgnome-l10nmessages-zhHK ... Done 066 066 Download: SUNWfirefoxl10n-it-IT ... Done 066 066 Download: SUNWgsf This is a headless machine so there is no monitor connected to it. I had just finished installing the machine and configuring the IP address so no one else was on the machine to reboot it had a clue of it's existence. I also kept having the same problem (machine rebooting with mkfile command) untill I tried the -k option on the kernel line. You are right, there is probably some hardware issue that is causing this (I am not blaming the OS), I am just trying to figure out what it is and how to find the cause. And no panic information in /var/adm/messages or on the console is very disturbing indeed. -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Quick Question. pkginfo -q equivalent in IPS
On Mon, Dec 14, 2009 at 12:02 AM, Shawn Walker swal...@opensolaris.org wrote: Peter Tribble wrote: On Thu, Dec 3, 2009 at 7:43 PM, Shawn Walker swal...@opensolaris.org wrote: Vikash Tulsiyan wrote: Do we have a pkginfo -q equivalent in IPS. I want to just check if a package is installed without all the details Check the return code of pkg info name; if it's 0, the package is installed, otherwise it isn't installed. Apart from the minor irritant of having to throw away the output, using pkg is significantly more expensive, particularly for packages that are installed: ptime pkginfo -q SUNWtcsh real 0.032076561 user 0.008303552 sys 0.017553991 ptime pkg info SUNWtcsh real 2.317786366 user 1.105758292 sys 0.876040367 ptime pkginfo -q froggle real 0.027556938 user 0.007866105 sys 0.016008880 ptime pkg info froggle real 0.649661522 user 0.359215141 sys 0.252858754 It's sufficiently slow that using pkg to work out whether packages are or aren't installed isn't really a viable option. And which version of pkg or build of OpenSolaris are you using to do that? 128a This is what I see: $ ptime pkg info SUNWtcsh pkg: info: no packages matching the following patterns you specified are installed on the system. Try specifying -r to query remotely: SUNWtcsh real 0.242818029 user 0.199911385 sys 0.042019782 ...but I'm running build 129. There have been major performance improvements in pkg in recent builds. And sorry, but you're complaining about .64 seconds? Seriously? Absolutely. Do that multiple times across hundreds of servers and it becomes considerable. The conclusion is that you;ll avoid using pkg to determine whether a package is installed, but end up using some other mechanism, like checking for a file you hope will always be in the package. -- -Peter Tribble http://www.petertribble.co.uk/ - http://ptribble.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Quick Question. pkginfo -q equivalent in IPS
On Thu, Dec 3, 2009 at 11:13 PM, Shawn Walker swal...@opensolaris.org wrote: Andre Lue wrote: Shawn, Can you say if there are any future plans to include a switch to override installing the dependencies of a package. No, there are no plans for that. If a package declares dependencies as required, then the package system must install them to be able to manage packages properly. To do so otherwise is madness from a dependency resolution standpoint. What's more important: the correct operation of the system, or the packaging system? With that said, it is planned that publication tools are enhanced so that you can easily copy a package and strip its dependencies if you want to customise it and install your own version. That way is madness from a systems administration standpoint. In terms of the operation of the packaged software, whether you override the packaging tools and force an install in violation of the dependencies, or whether you take another package and lie about its dependencies makes no difference - the same bits end up on the disk. In terms of manageability and problem resolution, there's a huge difference. In one case the packaging system can help the admin by telling you that there's a problem, and even let you clamber out of the hole you've landed in. In the other case you have a system that's broken and have no hope of diagnosing or fixing it. -- -Peter Tribble http://www.petertribble.co.uk/ - http://ptribble.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Quick Question. pkginfo -q equivalent in IPS
Peter Tribble wrote: On Mon, Dec 14, 2009 at 12:02 AM, Shawn Walker swal...@opensolaris.org wrote: And sorry, but you're complaining about .64 seconds? Seriously? Absolutely. Do that multiple times across hundreds of servers and it becomes considerable. The conclusion is that you;ll avoid using pkg to determine whether a package is installed, but end up using some other mechanism, like checking for a file you hope will always be in the package. If you can actually demonstrate that as an issue, you might be onto something. But trying to optimise for hypothetical situations doesn't seem beneficial. As I also mentioned before, you don't have to query for one package at a time. If you actually wanted to check the status of multiple packages, just use: pkg list ...and pipe the output to grep. That avoids the overhead of interpreter startup time, etc. -- Shawn Walker ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Quick Question. pkginfo -q equivalent in IPS
Peter Tribble wrote: On Thu, Dec 3, 2009 at 11:13 PM, Shawn Walker swal...@opensolaris.org wrote: Andre Lue wrote: Shawn, Can you say if there are any future plans to include a switch to override installing the dependencies of a package. No, there are no plans for that. If a package declares dependencies as required, then the package system must install them to be able to manage packages properly. To do so otherwise is madness from a dependency resolution standpoint. What's more important: the correct operation of the system, or the packaging system? The correct operation of the system, which is why the package system must also be correct ;) With that said, it is planned that publication tools are enhanced so that you can easily copy a package and strip its dependencies if you want to customise it and install your own version. That way is madness from a systems administration standpoint. That depends on who you ask. What's more mad, a system that may or may not operate correctly since the dependencies of the packages installed may or may not be satisfied, or one that is *known* to be correct? -- Shawn Walker ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Jumbo frames on rge
Not sure I've got too many answers, but suspiciously :-) I should say interestingly In : http://src.opensolaris.org/source/xref/onnv/onnv-gate/usr/src/uts/common/io/rge/rge_hw.h#RGE_BUFF_SIZE_JUMBO I see 542 /* 543 * Architectural constants: absolute maximum numbers of each type of ring 544 */ 545 546 #define RGE_SEND_SLOTS 1024 547 #define RGE_RECV_SLOTS 1024 548 #define RGE_BUFF_SIZE_STD 1536/* 1536 bytes */ 549 #define RGE_BUFF_SIZE_JUMBO 7168/* maximum 7K */ 550 #define RGE_JUMBO_SIZE 7014 551 #define RGE_JUMBO_MTU 7000 552 #define RGE_STATS_DUMP_SIZE 64 Maybe this is having some effect, or I might be off in the weeds. Now we need an rge driver expert ... .s -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] MP3 player client
Just finished writing a GUI client for MP3 players that use MTP (and not UMS) to upload/download files to/from a PC. Media Transfer Protocol (MTP) is commonly used by Microsoft, iRiver and most mobile phone manufacturers as the preferred method to connect MP3 players and phones to a PC. It's targeted at Solaris 10, so it's requirements are fairly low, and should run on OpenSolaris without problems. http://chewy509.110mb.com/gMTP.html For Solaris 10 and OpenSolaris, the only external dependency is libmtp, (other than having the development headers installed). Comments as always are welcomed. PS. UMS = USB Mass Storage. -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] ZFS Dedupe reporting incorrect savings
Hi, Created a zpool with 64k recordsize and enabled dedupe on it. zpool create -O recordsize=64k TestPool device1 zfs set dedup=on TestPool I copied files onto this pool over nfs from a windows client. Here is the output of zpool list Prompt:~# zpool list NAME SIZE ALLOC FREECAP DEDUP HEALTH ALTROOT TestPool 696G 19.1G 677G 2% 1.13x ONLINE - When I ran a dir /s command on the share from a windows client cmd, I see the file size as 51,193,782,290 bytes. The alloc size reported by zpool along with the DEDUP of 1.13x does not addup to 51,193,782,290 bytes. According to the DEDUP (Dedupe ratio) the amount of data copied is 21.58G (19.1G * 1.13) Here is the output from zdb -DD Prompt:~# zdb -DD TestPool DDT-sha256-zap-duplicate: 33536 entries, size 272 on disk, 140 in core DDT-sha256-zap-unique: 278241 entries, size 274 on disk, 142 in core DDT histogram (aggregated over all DDTs): bucket allocated referenced __ __ __ refcnt blocks LSIZE PSIZE DSIZE blocks LSIZE PSIZE DSIZE -- -- - - - -- - - - 1 272K 17.0G 17.0G 17.0G 272K 17.0G 17.0G 17.0G 232.7K 2.05G 2.05G 2.05G65.6K 4.10G 4.10G 4.10G 4 15960K960K960K 71 4.44M 4.44M 4.44M 84256K256K256K 53 3.31M 3.31M 3.31M 161 64K 64K 64K 16 1M 1M 1M 5121 64K 64K 64K 854 53.4M 53.4M 53.4M 1K1 64K 64K 64K1.08K 69.1M 69.1M 69.1M 4K1 64K 64K 64K5.33K341M341M341M Total 304K 19.0G 19.0G 19.0G 345K 21.5G 21.5G 21.5G dedup = 1.13, compress = 1.00, copies = 1.00, dedup * compress / copies = 1.13 Am I missing something? Your inputs are much appritiated. Thanks, Giri -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] ZFS Questions
I've been working trying to create a minimal opensolaris VirtualBox appliance and it looks like by default some sort of mirroring is enabled within the filesystem. Is it possible/easy to disable this. The appliance is just supposed to be a static development environment for the users and will be backed up separately anyway. Can you elaborate a bit on how you come to this conclusion? How do you see that there are mirroring in place? How did you install your OpenSolaris? Which version of OpenSolaris are you running? What version of VirtualBox? On what platform are you running Virtualbox? -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org