[opensuse-factory] Out of memory - endless loop
Gents,
I know that in certain circles, PHP has not a good
reputation( security) Untill know, i had never any problems.
I have a machine, not directly connected to the Net, and the only php
application on it is cacti.
Yesterday, coming home i found the machine unresponsive, it did reply to
ping, but that was all i could see (console dead, no reply to ssh)
After rebooting i found this in messages (below).
This section of messages repeated about hundred times
Is this a know feature of 10.1??
If there's a runaway process, too bad, that's life, it can happen.
But ending up in an endless loop is not good at all.
As it states free swap = 0kb it looks like a kernel process has eaten
all the mem, and that php gives the final trigger!
As it only happened once: i'll guess it's not reproducable.
Hans
==
nov 13 18:55:01 fw3 /usr/sbin/cron[963]: (wwwrun) CMD (
php /usr/share/cacti/poller.php /dev/null)
Nov 13 19:00:01 fw3 /usr/sbin/cron[998]: (wwwrun) CMD (
php /usr/share/cacti/poller.php /dev/null)
Nov 13 19:05:11 fw3 kernel: printk: 504 messages suppressed.
Nov 13 19:05:11 fw3 kernel: oom-killer: gfp_mask=0x201d2, order=0
Nov 13 19:05:11 fw3 kernel:
Nov 13 19:05:11 fw3 kernel: Call Trace: 8015ae11{out_of_memory
+53} 8015cc81{__alloc_pages+552}
Nov 13 19:05:11 fw3 kernel:
8015e4af{__do_page_cache_readahead+149}
802ceaf4{thread_return+0}
Nov 13 19:05:11 fw3 kernel:8015a477{filemap_nopage
+323} 80164b59{__handle_mm_fault+911}
Nov 13 19:05:11 fw3 kernel:802d1f7e{do_page_fault+965}
8018c9c7{sys_select+795}
Nov 13 19:05:11 fw3 kernel:8010b449{error_exit+0}
Nov 13 19:05:11 fw3 kernel: Mem-info:
Nov 13 19:05:11 fw3 kernel: Node 0 DMA per-cpu:
Nov 13 19:05:11 fw3 kernel: cpu 0 hot: high 0, batch 1 used:0
Nov 13 19:05:11 fw3 kernel: cpu 0 cold: high 0, batch 1 used:0
Nov 13 19:05:12 fw3 kernel: cpu 1 hot: high 0, batch 1 used:0
Nov 13 19:05:13 fw3 kernel: cpu 1 cold: high 0, batch 1 used:0
Nov 13 19:05:13 fw3 kernel: Node 0 DMA32 per-cpu:
Nov 13 19:05:13 fw3 kernel: cpu 0 hot: high 186, batch 31 used:20
Nov 13 19:05:13 fw3 kernel: cpu 0 cold: high 62, batch 15 used:24
Nov 13 19:05:14 fw3 kernel: cpu 1 hot: high 186, batch 31 used:8
Nov 13 19:05:14 fw3 kernel: cpu 1 cold: high 62, batch 15 used:61
Nov 13 19:05:14 fw3 kernel: Node 0 Normal per-cpu: empty
Nov 13 19:05:14 fw3 kernel: Node 0 HighMem per-cpu: empty
Nov 13 19:05:14 fw3 kernel: Free pages:8016kB (0kB HighMem)
Nov 13 19:05:14 fw3 kernel: Active:104983 inactive:112014 dirty:0
writeback:0 unstable:0 free:2004 slab:13709 mapped:204535
pagetables:15090
Nov 13 19:05:14 fw3 kernel: Node 0 DMA free:4016kB min:44kB low:52kB
high:64kB active:3400kB inactive:3208kB present:12056kB
pages_scanned:16107 all_unreclaimable? yes
Nov 13 19:05:14 fw3 kernel: lowmem_reserve[]: 0 993 993 993
Nov 13 19:05:15 fw3 kernel: Node 0 DMA32 free:4000kB min:4008kB
low:5008kB high:6012kB active:416532kB inactive:444848kB
present:1017828kB pages_scanned:2416680 all_unreclaimable? yes
Nov 13 19:05:15 fw3 kernel: lowmem_reserve[]: 0 0 0 0
Nov 13 19:05:15 fw3 kernel: Node 0 Normal free:0kB min:0kB low:0kB
high:0kB active:0kB inactive:0kB present:0kB pages_scanned:0
all_unreclaimable? no
Nov 13 19:05:15 fw3 /usr/sbin/cron[1069]: (wwwrun) CMD (
php /usr/share/cacti/poller.php /dev/null)
Nov 13 19:05:15 fw3 kernel: lowmem_reserve[]: 0 0 0 0
Nov 13 19:05:15 fw3 kernel: Node 0 HighMem free:0kB min:128kB low:128kB
high:128kB active:0kB inactive:0kB present:0kB pages_scanned:0
all_unreclaimable? no
Nov 13 19:05:15 fw3 kernel: lowmem_reserve[]: 0 0 0 0
Nov 13 19:05:15 fw3 kernel: Node 0 DMA: 0*4kB 0*8kB 1*16kB 1*32kB 0*64kB
1*128kB 1*256kB 1*512kB 1*1024kB 1*2048kB 0*4096kB = 4016kB
Nov 13 19:05:15 fw3 kernel: Node 0 DMA32: 0*4kB 2*8kB 7*16kB 1*32kB
0*64kB 4*128kB 1*256kB 2*512kB 0*1024kB 1*2048kB 0*4096kB = 4000kB
Nov 13 19:05:15 fw3 kernel: Node 0 Normal: empty
Nov 13 19:05:15 fw3 kernel: Node 0 HighMem: empty
Nov 13 19:05:15 fw3 kernel: Swap cache: add 420648, delete 420648, find
82842/99473, race 0+39
Nov 13 19:05:16 fw3 kernel: Free swap = 0kB
Nov 13 19:05:16 fw3 kernel: Total swap = 1052248kB
Nov 13 19:05:16 fw3 kernel: Free swap:0kB
Nov 13 19:05:16 fw3 kernel: 262080 pages of RAM
Nov 13 19:05:16 fw3 kernel: 5020 reserved pages
Nov 13 19:05:16 fw3 kernel: 9462 pages shared
Nov 13 19:05:16 fw3 kernel: 0 pages swap cached
Nov 13 19:05:16 fw3 kernel: Out of Memory: Kill process 32061 (php)
score 62063 and children.
Nov 13 19:05:16 fw3 kernel: Out of memory: Killed process 32065 (php).
==
--
pgp-id: 926EBB12
pgp-fingerprint: BE97 1CBF FAC4 236C 4A73 F76E EDFC D032 926E BB12
Registered linux user: 75761 (http://counter.li.org)
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For
Re: [opensuse-factory] Out of memory - endless loop
Hans Witvliet wrote:
Gents,
I know that in certain circles, PHP has not a good
reputation( security) Untill know, i had never any problems.
PHP has problems, but the biggest problem is the code quality of the
applications written on it.
Nov 13 19:05:11 fw3 kernel: Call Trace: 8015ae11{out_of_memory
+53} 8015cc81{__alloc_pages+552}
This is a kernel crash that is most likely caused by something that ahs
nothing to do with PHP.
How can I reproduce it ? ( step By step ? )
signature.asc
Description: OpenPGP digital signature
Re: [opensuse-factory] lua50 not synced out
Andreas Hanke [EMAIL PROTECTED] writes: Hi, I have a problem with the following two bugs: https://bugzilla.novell.com/show_bug.cgi?id=217875 https://bugzilla.novell.com/show_bug.cgi?id=219773 The former is not properly fixed and needs to be reopened, but the latter prevents me from working on a solution. Would it be possible to get lua50 synced out on the next opportunity? It's missing since it exists (~4 weeks according to my crystal ball). We're looking into this now, thanks for the report, Andreas -- Andreas Jaeger, [EMAIL PROTECTED], http://www.suse.de/~aj/ SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126 pgpF6fgBSLQCA.pgp Description: PGP signature
Re: [opensuse-factory] VMWare-Config [was: Linux-Kernel-Headers outdated]
Reply on 14-11-2006 13:00:22 I installed another 10.1 for this test. VMware-tools installs there, no problem with the kernel modules. Anyway, what I wanted to see was if the tools would work any better in 10.2 than the have so far in Linux. The windows client can run in its window and the cursor can go in and out of the window when tools are installed. Can you confirm this function working when running SUSE as guest? This has not been the case here. The cursor needs to be released from the guest window by pressing Ctrl+Alt if the tools are there or not. So the tools in case of SUSE as guest are useless anyway. The display properties can be adjusted just as fine with or without the tools, so I'm not too worried about the tools. Also if I run a remote session to that guest SUSE with FreeNX, the window works as any other window. Tools = no biggie :) No, I can not confirm this. This still seems not to work with Linux Guests (I don't think it's specific to SUSE). So the tools are really limited in usability... Also the clipboard copy/paste functionaltity, which would be VERY handy, seems not to work (VM WS 5.5 on Win Host) Dominique
Re: [opensuse-factory] How to name that 'thing' ?
Reply on 14-11-2006 15:16:24 Hi, looking at http://en.opensuse.org/Package_Management#Package_source_directories_on_t he_servers it seems that there is no agreement on how to name a 'software repository', 'installation source', 'catalog', 'channel', whatever ;-) I'm hereby asking the community to come forward with a proposal for a proper name. Klaus, I think the problem arised especially from the different tools to manage these 'things' yum - repositories smart - channels yast - installation sources rug/zmd - catalogs (But I'm sure you knew this) I for myself always use the wording 'repository', which is also very reflected on the BuildService (repos.opensuse.org). But probably just a usability question. But I agree, it would be nice to have these wordings unified, but again: this will be VERY difficult when you explain the users how to use a specific program and they come with specific problems. Dominique
[opensuse-factory] Grub Bugs
Tha last week, I have made a installation in a new har disk with 10.2 beta 1, and later upgraded to beta 2. Later I has cloned the hard disk, and when I boot from the 1st cd of beta 1 to correct and repair the grub installation, I could'nt add other OS residing in other hard disk, and the yast/grub utilitý do'nt let me add the disk and partition from the other OS. Later, I tryied to repair grub with the 1st Cd of SuseLinux 10.1(from the box), and the config utility works ok, but it coul'nt write the changes to the disks. I made a search in bugzila with the grub keyword, and I found 21 with the status NEW, about 60 ASSIGNED, and 19 with NEEDINFO status. https://bugzilla.novell.com/buglist.cgi?quicksearch=grubx=17y=4 How can I deal with this problem? Registering a new bug, I mean is'nt a solution Thanks - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse-factory] Linux-Kernel-Headers outdated
Vahis wrote: Dominique Leuenberger wrote: Reply on 13-11-2006 16:00:21 Dominique Leuenberger wrote: Hi everybody, I just posted bugzilla https://bugzilla.novell.com/show_bug.cgi?id=220384 Just installed Beta 2 in a VMware. In order to compile the Tools, the Kernel Header include files are quired. Installed the package 'linux-kernel-headers', but they are from an older kernel (2.6.18-13, while installed kernel is 2.6.18.2-4) Also in factroy tree, linux-kernel-headers is at 2.6.18-13 Severity is set to normal at the moment, even though it should be much higher (but I assume as reporter I should not change that) OR: Did I miss something completely different? (that would render that bugzilla report invalid? Regards, Dominique - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] I have seen this vmware-tools/kernel header not matching -problem discussed in several groups and forums concerning several distros in addition to SUSE. And not just rpm based ones. As for SUSE it has been discussed already during the 9.x series. But I have still not found a solutions for it. I sure would like to see a solution how to install the vmwaretools properly :) Although this is obviously not quite the right place, I'd like to use the opportunity to ask for any pointers to this matter. Vahis, please read the Bugzilla Followup. I'm sorry for the noise I did. But it seems it's not the package linux-kernel-headers that is required, but the whole kernel-source. After installing that package, I was able to run vmware-config-tools.pl without any problems. Regards, Dominique I installed another 10.1 for this test. VMware-tools installs there, no problem with the kernel modules. Anyway, what I wanted to see was if the tools would work any better in 10.2 than the have so far in Linux. The windows client can run in its window and the cursor can go in and out of the window when tools are installed. Can you confirm this function working when running SUSE as guest? This has not been the case here. The cursor needs to be released from the guest window by pressing Ctrl+Alt if the tools are there or not. So the tools in case of SUSE as guest are useless anyway. The display properties can be adjusted just as fine with or without the tools, so I'm not too worried about the tools. Also if I run a remote session to that guest SUSE with FreeNX, the window works as any other window. Tools = no biggie :) Nevertheless the tools, but I came to think why it perhaps does not compile. I installed a real vmware server on a real host machine running 10.2 beta2. There weren't any probs. I haven't installed any virtual machines on that yet, but I guess the server should be like the tools in that matter. So I think maybe this problem with the vmware-tools not compiling comes from the fact that my guest machine was built with the http://www.easyvmx.com/ service. It's fake It's not installed from a real vmware workstation like they used to be when I made the instructions on my site. http://waxborg.pp.fi/mobile/articles/vmware.html Anything to confirm this? -- Vahis Sometimes I reply to top posters. Seldom. And usually just once. Motorcycling, Boating and SUSE Linux: http://waxborg.servepics.com - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse-factory] Software test plans
On 11/14/06, JP Rosevear [EMAIL PROTECTED] wrote: bugzilla.novell.com now has testopia 1.1 as well and there is some testing going on with that, could be a route forward. Yes, I think that Testopia will be very useful to allow community beta testers to contribute further to the formal QA test cycle, and as I learn more I will document on http://en.opensuse.org/testing As some of the emails to this thread have shown, there are people just sitting at home wishing to help, they just need some guidance. And I think there is room for the casual tester who just wants a quick test to run through. I do not really know enough yet, but I think Testopia may be a bit heavy for those users, and may provide an initial hurdle. So I can see Wiki based testplans for new and casual users to be more of a guide than formal test plans and then if testers want to give more time they can graduate to actually doing official test runs of the Testopia testplans and record results against them in the bugzilla. Pflodo Peter Flodin - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
[opensuse-factory] How to name that 'thing' ?
Hi, looking at http://en.opensuse.org/Package_Management#Package_source_directories_on_the_servers it seems that there is no agreement on how to name a 'software repository', 'installation source', 'catalog', 'channel', whatever ;-) I'm hereby asking the community to come forward with a proposal for a proper name. Klaus - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse-factory] Proposal on metadata fetching
Reply on 14-11-2006 15:34:53 For a detailed search do not download metadata but query a server. E.g if im looking for the string "foobar" and content cannot be found in the local metadatafile, Suse should not download further metadata, but contact an official server, which does the search. I do not know, what kind of server we need, if 10 or 100 thousends of clients will start search queries, but I think this may be faster, than fetching so much metadata all the time. I also do not know how many data must be uploaded in order to get this system working, so maybe it's a very bad idea and not realizeable, but perhaps there is a chance. Marcel, The idea of only having the RPM names / Summary is not bad IMHO. (maybe having an option for the user to decide HOW he would like to add the repo?) The thing with the search engine sounds fascinating, but I don't think it's possible to realize: just think about all the repos that exist out there. How could they possibly EVER be included in a search? And in plus, the server you query should know in which repos to have a look, as obviously it should only offer packages from repos, you're subscribed to. Dominique
Re: [opensuse-factory] Proposal on metadata fetching
Marcel Hilzinger schrieb: The big difference between Suse and Debian/Ubuntu is, that under Suse much more data get's fetched. So you can also search within package-descriptions etc, whereas under Ubuntu you have only package-names and files, but no description e.g. For a more detailed search you need apt-file. I think apt-file is compearable to the actual suse solution. As a starting point, the YaST2 metadata could be compressed. Debian does that since ever, but YaST2 metadata were never compressed. The packages file in Factory is currently at 16 MB, a good bzip2 compression reduces it to 1.8 MB - almost 90% gone. I think that this should really be considered first because it is possible entirely without functionality loss and without complicating the architecture too much. The thing with the patterns for unneeded architectures is more a cosmetic thing because the pattern files are very small. But many users notice it because YaST displays the filenames of the downloaded metadata. For YUM metadata, there is not much to do because they are already compressed and it's a standardized format - it can be extended, but not changed in incompatible ways. Andreas - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
[opensuse-factory] Proposal on metadata fetching
Beta2 is much better than 10.1 concerning the speed of the package manager. But as mentioned by other users already it's still far behind debian concerning the speed of fetching metadata. Adding factory as an installation source takes at least 15 minutes with an average DSL connection. The big difference between Suse and Debian/Ubuntu is, that under Suse much more data get's fetched. So you can also search within package-descriptions etc, whereas under Ubuntu you have only package-names and files, but no description e.g. For a more detailed search you need apt-file. I think apt-file is compearable to the actual suse solution. My proposal (for 10.3 etc): Reduce downloaded metadata to package-names and RPM-content. Do not fetch descriptions, etc. Most users know the name of the file/RMP they want to install. An error tolerant search can help finding the right package. For a detailed search do not download metadata but query a server. E.g if im looking for the string foobar and content cannot be found in the local metadatafile, Suse should not download further metadata, but contact an official server, which does the search. I do not know, what kind of server we need, if 10 or 100 thousends of clients will start search queries, but I think this may be faster, than fetching so much metadata all the time. I also do not know how many data must be uploaded in order to get this system working, so maybe it's a very bad idea and not realizeable, but perhaps there is a chance. -- Mit freundlichen Grüßen, Marcel Hilzinger Linux New Media AG Süskindstr. 4 D-81929 München Tel: +49 (89) 99 34 11 0 Fax: +49 (89) 99 34 11 99 - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse-factory] How to name that 'thing' ?
James Ogley [EMAIL PROTECTED] writes: You mean you don't want a big pile of unmaintainable patches changing every instance of 'channel' to 'repository'? Why-ever not? ;) [...] I add mine, +1 for repository. If you are looking for something new, try (software) directory. -- Karl Eichwalder RD / DocumentationSUSE Linux Products GmbH GPG: 1024D/06EB882E B2A3 AF2F CFC8 40B1 67EA 475A 5903 A21B 06EB 882E - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse-factory] How to name that 'thing' ?
Robert Schiele [EMAIL PROTECTED] writes: On Tue, Nov 14, 2006 at 03:19:19PM +0100, Karl Eichwalder wrote: James Ogley [EMAIL PROTECTED] writes: You mean you don't want a big pile of unmaintainable patches changing every instance of 'channel' to 'repository'? Why-ever not? ;) [...] I add mine, +1 for repository. If you are looking for something new, try (software) directory. Or to be even more innovative: Make this a configure option in YaST. Yeah, but let's avoid the term source (as in installation source)--that's too closely coupled with source code. -- Karl Eichwalder RD / DocumentationSUSE Linux Products GmbH GPG: 1024D/06EB882E B2A3 AF2F CFC8 40B1 67EA 475A 5903 A21B 06EB 882E - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse-factory] How to name that 'thing' ?
* Dominique Leuenberger [EMAIL PROTECTED] [Nov 14. 2006 14:19]: I think the problem arised especially from the different tools to manage these 'things' yum - repositories smart - channels yast - installation sources rug/zmd - catalogs (But I'm sure you knew this) ;-) We cannot change tools developed outside of the opeSUSE community. But we do want to unify the wording for tools we develop like YaST, zypper, opensuse-updater, etc. I for myself always use the wording 'repository', which is also very reflected on the BuildService (repos.opensuse.org). I counted your vote ;-) Klaus - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse-factory] How to name that 'thing' ?
We cannot change tools developed outside of the opeSUSE community. You mean you don't want a big pile of unmaintainable patches changing every instance of 'channel' to 'repository'? Why-ever not? ;) I for myself always use the wording 'repository', which is also very reflected on the BuildService (repos.opensuse.org). I counted your vote ;-) I add mine, +1 for repository. -- James Ogley [EMAIL PROTECTED] http://usr-local-bin.org Packages for SUSE: http://usr-local-bin.org/rpms Help end poverty: http://oxfam.org.uk/imin - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
[opensuse-factory] USB keyfob and sync
I haven't yet braved 10.2 (I'm using 10.1 for daily desktop work), but I am curious: On the last few versions, including 10.1, I have had to create a file /usr/share/hal/fdi/policy/20thirdparty/90-nosync.fdi which contains: ?xml version=1.0 encoding=UTF-8? deviceinfo version=0.2 device !-- disable sync for mount -- match key=block.is_volume bool=true match key=volume.fsusage string=filesystem match key=@info.parent:storage.bus string=usb merge key=volume.policy.mount_option.sync type=boolfalse/merge /match /match /match /device /deviceinfo The purpose of course is to turn off sync on usb keyfobs because of the terrible throughput rate that results if sync is on. Has this been fixed for 10.2? Or will I need to create this file again if I upgrade? --Jon Robison
Re: [opensuse-factory] amarok-helix
Marcel Hilzinger [EMAIL PROTECTED] writes: Has it been dropped totally? Under inst-source-nonoss there are only helix-banshee packages, but no amarok. Yes, we needed to reorganise our complete packaging for banshee and amarok. We can only use the real codecs from Real - or Helix - packages like Helix Banshee, Andreas -- Andreas Jaeger, [EMAIL PROTECTED], http://www.suse.de/~aj/ SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126 pgpAyYGsMqlOj.pgp Description: PGP signature
Re: [opensuse-factory] Gaim 2.0
On 2006-11-14 16:14:20 +0100, Simon Strandman wrote: I want to suggest that gaim 2.0 is imported into factory beforce the RC. I like it much more than 1.5 which is over a year old anyway (released in august 2005). The current version is 2.0-beta5 but it's already stable and the final version is soon to be released. Fedora 6 also ships beta4. latest rumors say the 2.0betas are buggy as usual. spellchecking is said to be broken e.g. even for beta5. darix -- openSUSE - SUSE Linux is my linux openSUSE is good for you www.opensuse.org - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse-factory] ok box
Lukas Ocilka a écrit : So, here are some rules: [ OK ] [ Cancel ] [ OK ] [ Cancel ] [ Skip ] [ Abort ] [ Retry ] [ Abort ] [ Retry ] [ Ignore ] [ Yes ] [ No ] [ Continue ] [ Cancel ] abort and cancel are nearly the same, so * why two words? * why a different layout (abort on left, cancel on right)? thanks better have all this written. Consistency in the interface is of course not the first thing to do, but probably should be the second :-) jdd -- http://www.dodin.net http://dodin.org/mediawiki/index.php/GPS_Lowrance_GO - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse-factory] ok box
Andreas Jaeger wrote: jdd [EMAIL PROTECTED] writes: Andreas Jaeger a écrit : jdd [EMAIL PROTECTED] writes: I'm not sure of that, But I have the feeling than the boxes asking for OK, cancel, retry are not all the same. I mean that some times the OK button is on the right, sometimes on the left of the box (sometimes the OK is a big left icon, when most of the time the icons are only decoration. If those are in YaST: Please make screenshots and open a bugreport, well... what is the normal way? (not necessary to fill a bug report in that case) First case: cancel retry OK or second case OK cancel retry or??? I don't remember :-(. Jiri? Actually, we have some non-written rules. Or they are written but internally. Our plan is to release a wiki-based YaST Dialog Style Guide but after openSUSE 10.2 is out. So, here are some rules: [ OK ] [ Cancel ] [ OK ] [ Cancel ] [ Skip ] [ Abort ] [ Retry ] [ Abort ] [ Retry ] [ Ignore ] [ Yes ] [ No ] [ Continue ] [ Cancel ] Because these rules are still not officially 'written', it might be a bit hard to file a bug against them :) ;) Lukas signature.asc Description: OpenPGP digital signature
[opensuse-factory] VIA GPU vs. oSL10.2B2
Hello, I have a mainboard based GPU (01:00.0 VGA compatible controller: VIA Technologies, Inc. S3 Unichrome Pro VGA Adapter (rev 01) ) and the openSuSE Liux 10.2 Beta2 (x86_64) does not indentitify corectly it. The frame buffer mode does not work, the sreen scrolling is brakeing - i dont say better in english, sorry... Why do not setup via module (or, would be better, module for unicrome) the Sax2? Balazs - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse-factory] Linux-Kernel-Headers outdated
Dominique Leuenberger wrote: Reply on 13-11-2006 16:00:21 Dominique Leuenberger wrote: Hi everybody, I just posted bugzilla https://bugzilla.novell.com/show_bug.cgi?id=220384 Just installed Beta 2 in a VMware. In order to compile the Tools, the Kernel Header include files are quired. Installed the package 'linux-kernel-headers', but they are from an older kernel (2.6.18-13, while installed kernel is 2.6.18.2-4) Also in factroy tree, linux-kernel-headers is at 2.6.18-13 Severity is set to normal at the moment, even though it should be much higher (but I assume as reporter I should not change that) OR: Did I miss something completely different? (that would render that bugzilla report invalid? Regards, Dominique - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] I have seen this vmware-tools/kernel header not matching -problem discussed in several groups and forums concerning several distros in addition to SUSE. And not just rpm based ones. As for SUSE it has been discussed already during the 9.x series. But I have still not found a solutions for it. I sure would like to see a solution how to install the vmwaretools properly :) Although this is obviously not quite the right place, I'd like to use the opportunity to ask for any pointers to this matter. Vahis, please read the Bugzilla Followup. I'm sorry for the noise I did. But it seems it's not the package linux-kernel-headers that is required, but the whole kernel-source. After installing that package, I was able to run vmware-config-tools.pl without any problems. Regards, Dominique I installed another 10.1 for this test. VMware-tools installs there, no problem with the kernel modules. Anyway, what I wanted to see was if the tools would work any better in 10.2 than the have so far in Linux. The windows client can run in its window and the cursor can go in and out of the window when tools are installed. Can you confirm this function working when running SUSE as guest? This has not been the case here. The cursor needs to be released from the guest window by pressing Ctrl+Alt if the tools are there or not. So the tools in case of SUSE as guest are useless anyway. The display properties can be adjusted just as fine with or without the tools, so I'm not too worried about the tools. Also if I run a remote session to that guest SUSE with FreeNX, the window works as any other window. Tools = no biggie :) -- Vahis Sometimes I reply to top posters. Seldom. And usually just once. Motorcycling, Boating and SUSE Linux: http://waxborg.servepics.com - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse-factory] ok box
Am Tuesday 14 November 2006 17:57 schrieb jdd: Lukas Ocilka a écrit : So, here are some rules: [ OK ] [ Cancel ] [ OK ] [ Cancel ] [ Skip ] [ Abort ] [ Retry ] [ Abort ] [ Retry ] [ Ignore ] [ Yes ] [ No ] [ Continue ] [ Cancel ] abort and cancel are nearly the same, so * why two words? * why a different layout (abort on left, cancel on right)? it cames from the time, where Abort did close the complete YaST window and Cancel just canceld one action (esp. in a popup box). -- Adrian Schroeter SUSE Linux Products GmbH, Maxfeldstr. 5, 90409 Nuernberg, Germany email: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse-factory] How to name that 'thing' ?
Em Ter, 2006-11-14 às 14:19 +0100, Dominique Leuenberger escreveu:
smart - channels
Smart calls them channels in order to have a single name to avoid
confusion among the various ways of fetching/storing a file (not only
RPM).
I agree that for RPM the calling it 'repo' is much easier (and makes
sense since rpm-md uses a 'repomd' file and 'repodata' dir).
--
% Mauricio Teixeira (netmask)
% mteixeira{a}webset{d}net Maceio/AL/BR
% http://mteixeira.webset.net http://pmping.sf.net
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
[opensuse-factory] Software test plans
Forgot to send to this list. Sorry for that. On 11/14/06, Peter Flodin [EMAIL PROTECTED] wrote: On 11/14/06, JP Rosevear [EMAIL PROTECTED] wrote: bugzilla.novell.com now has testopia 1.1 as well and there is some testing going on with that, could be a route forward. Yes, I think that Testopia will be very useful to allow community beta testers to contribute further to the formal QA test cycle, and as I learn more I will document on http://en.opensuse.org/testing As some of the emails to this thread have shown, there are people just sitting at home wishing to help, they just need some guidance. And I think there is room for the casual tester who just wants a quick test to run through. I do not really know enough yet, but I think Testopia may be a bit heavy for those users, and may provide an initial hurdle. Why not start and test SyncML and OBEX. Very nice guide: http://en.opensuse.org/OpenSync/SyncML-OBEX-Client Warm Regards, Claes - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse-factory] amarok-helix
Tirsdag 14 november 2006 16:18 skrev Andreas Jaeger: Yes, we needed to reorganise our complete packaging for banshee and amarok. We can only use the real codecs from Real - or Helix - packages like Helix Banshee, Can as in Real will only allow.. or as in technical necessity? Martin - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse-factory] amarok-helix
Martin Schlander schrieb: Can as in Real will only allow.. or as in technical necessity? Look at banshee vs. helix-banshee and you'll see the reason. Without knowing the exact background information, I guess that SUSE would have been forced to rename Amarok to Helix Amarok, add a Helix EULA on startup and add Helix branding. The changelog of the amarok package cleary states that amarok-helix has been disabled on purpose. Certainly not in order to annoy the users. Andreas Hanke - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse-factory] amarok-helix
Andreas Hanke [EMAIL PROTECTED] writes: Martin Schlander schrieb: Can as in Real will only allow.. or as in technical necessity? Look at banshee vs. helix-banshee and you'll see the reason. Without knowing the exact background information, I guess that SUSE would have been forced to rename Amarok to Helix Amarok, add a Helix EULA on startup and add Helix branding. And we could not do so since we do not own the Amarok copyright - but did own the banshee copyright... Andreas -- Andreas Jaeger, [EMAIL PROTECTED], http://www.suse.de/~aj/ SUSE Linux Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126 pgpxZoce8X0jD.pgp Description: PGP signature
Re: [opensuse-factory] Font Funkiness
Randall R Schulz [EMAIL PROTECTED] さんは書きました: Hi, I just witnessed an odd phenomenon and I'm wondering if it's a sign of a problem or an aspect of the way font rendering operates (I can imagine it either way). Specifically, I opened the Beagle window by clicking its tray icon. When the window appears, the fonts were extremely ragged. I decided to take a snapshot, so I used the system (KDE / Geeko) menu to launch KSnapshot. I took a picture of the Beagle window. While I was doing so, I was paying attention only to the KSnapshot interface and had neglected to really watch the Beagle window. When I finally dismissed the KSnapshot window (after capturing and saving the picture) I was startled to see that the ragged font display had been replaced with a properly smoothed one. I've seen the phenomenon repeat itself after this first instance (again in the Beagle GUI). So my question is this: Is this a bug or is there some kind of asynchronous rendering of smoothed fonts that sometimes leads to a ragged font display upon first use of a given font but which is later replaced with the smoothed version, once the smoothed glyphs have been rendered? This might be the following bug: http://bugzilla.novell.com/show_bug.cgi?id=193095 -- Mike FABIAN [EMAIL PROTECTED] http://www.suse.de/~mfabian 睡眠不足はいい仕事の敵だ。 I � Unicode - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse-factory] How to name that 'thing' ?
Klaus Kaempf wrote: Hi, looking at http://en.opensuse.org/Package_Management#Package_source_directories_on_the_servers it seems that there is no agreement on how to name a 'software repository', 'installation source', 'catalog', 'channel', whatever ;-) I'm hereby asking the community to come forward with a proposal for a proper name. Klaus - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] I'd vote for repository. Regards Sid. -- Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support Specialist, Cricket Coach Microsoft Windows Free Zone - Linux used for all Computing Tasks - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse-factory] Deselecting Zenworks
Hello,
Am Sonntag, 12. November 2006 12:34 schrieb Andreas Jaeger:
Keith Goggin [EMAIL PROTECTED] writes:
I found deselecting Zenworks and selecting instead the lightweight
alternative to Zenworks did not prevent the ubiquitous Zenworks
from being installed. Note the red icon along side the deselected
Zenworks is defined as Taboo-Never Install
I know this has already been reported but I think this is
potentially very serious at this late stage of testing.
There're already bug reports for these -
The only one I found (searching for pattern taboo) is
https://bugzilla.novell.com/show_bug.cgi?id=215355
(current state is INVALID...)
The problem is that marking a pattern as taboo does not mark the
contained packages as taboo. The packages are only unselected and can
be re-selected via dependencies (which happens with zmd).
This is not the expected behaviour and can cause confusion - especially
for the zenworks pattern which is probably the top taboo pattern ;-)
I'll check this next week,
Andreas, are there any results on this yet?
Regards,
Christian Boltz
--
vi-Befehle sind sogar relativ einfach zu merken. Wenn man einmal weiß,
was dw db de d) d( d} d{ dd d^ d$ d0 dG sowie cw und yw machen, dann
weiß man auch, was cb ce c) c( c} c{ cc c^ c$ c0 cG sowie yb ye y) y( y}
y{ yy y^ y$ y0 yG machen.[Bernd Bordesser in suse-linux]
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse-factory] Font Funkiness
Randall R Schulz [EMAIL PROTECTED] さんは書きました: On Tuesday 14 November 2006 13:11, Mike FABIAN wrote: ... So my question is this: Is this a bug or is there some kind of asynchronous rendering of smoothed fonts that sometimes leads to a ragged font display upon first use of a given font but which is later replaced with the smoothed version, once the smoothed glyphs have been rendered? This might be the following bug: http://bugzilla.novell.com/show_bug.cgi?id=193095 There doesn't appear to be any mention in that bug report of spontaneous replacement of the raggedly rendered glyphs with smoothed ones shortly after the original, ragged display. The bug report does mention this, see for example comment #23 Comment #23 From Felix Möller 2006-11-10 16:35:53 MST i just ran gnome-theme-manager as was said in 214280. This crashedfirefox and thunderbird. But now all font problems are gone! So just starting gnome-theme-manager and exiting it without doing anything solves this problem. Federico Mena Quintero recently explained me that GNOME works as follows: - The GNOME control center sets some GConf keys. - The running gnome-settings-daemon gets notified about changes in the GConf keys, and it does the following. First, it sets some X resources. Second, it uses the XSETTINGS protocol to tell GTK+ apps to refresh themselves. - Running GTK+ apps pick up the XSETTINGS notification, call cairo_set_font_options(), and redraw their widgets. You can try that out by starting for example gedit and then gnome-font-properties and change the anti-alias setting in gnome-font-properties. You will see that it applies to gedit immediately. I am not sure which GNOME programs apart from gnome-font-properties and gnome-theme-manager can trigger this process to make GTK+ apps to refresh themselves according to the current settings of the Xft.* resources. But probably there are others. Therefore I guess that some gnome thing started and triggered this process while you were trying to make the snapshot of Beagle (which is a GTK+ program). I assume rendered glyphs are cached (lest font rendering consume inordinate CPU cycles throughout the system's operation). To your knowledge, is there any code that might account for ragged (point-resampled) glyphs being used (transiently) when the smoothed ones are not available on a timely basis? No. -- Mike FABIAN [EMAIL PROTECTED] http://www.suse.de/~mfabian 睡眠不足はいい仕事の敵だ。 I � Unicode - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse-factory] VIA GPU vs. oSL10.2B2
You can try with the propietary drivers from VIA: http://www.viaarena.com/default.aspx?PageID=2OSID=19CatID=2170 2006/11/14, Kerti Balázs Gábor [EMAIL PROTECTED]: Hello, I have a mainboard based GPU (01:00.0 VGA compatible controller: VIA Technologies, Inc. S3 Unichrome Pro VGA Adapter (rev 01) ) and the openSuSE Liux 10.2 Beta2 (x86_64) does not indentitify corectly it. The frame buffer mode does not work, the sreen scrolling is brakeing - i dont say better in english, sorry... Why do not setup via module (or, would be better, module for unicrome) the Sax2? Balazs - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
[opensuse-factory] reboot instead of load kernel/initrd?
I'm really scratching my head over this. I have a system with HD on hda hdc, with 23 partitions on hda and 14 partitions on hdc. Among them I have installations of SUSE Factory, SUSE 10.0, Fedora 6, Debian Etch, Xandros 3, Ubuntu 6.10, Knoppix, DOS, OS/2 windoz. I have separate grub /boot partitions for Etch on hda5 and for SUSE 10.0 on hdc5. Each of the others have their own grub versions installed on their respective roots, except for Xandros, which has lilo on its root. The MBR of each HD has generic code. IBM Boot Manager is installed on hda2 and hdc2. I did some disk space reallocation so that some of the distros would have larger root partitions. So far I've only successfully done the move with 1 of 4-5 total planned. ATM, I've succeeded in transplanting Fedora from hda22 to hda11. I can boot hda11 either by chainloading from elsewhere, or by selecting it from IBM BM. I've succeeded in copying SUSE Factory from hda21 to hda10, but from hda10 it will only boot if I use one of the other distro's grubs to start it, either from a menu stanza, or directly from a grub prompt. From hda21 it still works as always. When I try to load Factory on hda10 with IBM BM or by chainloading to Factory's installed grub, the system instantly cold boots. I've tried redoing the copy from hda21 to hda10 twice (total 3, with fresh mkfs -t ext3 each time), and tried grub-install by chrooting from a different distro's boot, by booting it, and by installing with --root-directory from a 10.0 boot, always with no errors reported, but nothing changes anything. Grub reports no errors, so I have no idea how to troubleshoot. Any ideas? -- Rejoice and be glad, because great is your reward in heaven. Matthew 5:12 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://mrmazda.no-ip.com/ - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse-factory] Proposal on metadata fetching
On Tuesday 14 November 2006 07:46, Andreas Hanke wrote: Marcel Hilzinger schrieb: The big difference between Suse and Debian/Ubuntu is, that under Suse much more data get's fetched. So you can also search within package-descriptions etc, whereas under Ubuntu you have only package-names and files, but no description e.g. For a more detailed search you need apt-file. I think apt-file is compearable to the actual suse solution. As a starting point, the YaST2 metadata could be compressed. Debian does that since ever, but YaST2 metadata were never compressed. The packages file in Factory is currently at 16 MB, a good bzip2 compression reduces it to 1.8 MB - almost 90% gone. I think that this should really be considered first because it is possible entirely without functionality loss and without complicating the architecture too much. ... Besides compression of metadata, the idea of central server that will contain all details of repositories is appealing because that will substantially simplify package management from user prospective. One preconfigured server to query for detailed information. The problem is, who will host such server, if even link to some repositories can be source of legal trouble. -- Regards, Rajko M. - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse-factory] Problem em canceling aspect ratio configuration
Jim Pye wrote: On Mon, 2006-11-13 at 22:02 +, Hugo Costelha wrote: I had the same problem when testing the configuration. The bug report is here https://bugzilla.novell.com/show_bug.cgi?id=216451 Thanks for trying this out. Hugo Costelha - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] Hugo Added my vote and 00.02 to this. Even though the bug has a resolved status I do not think it is working properly. So hopefully will be reopened and looked at. However it's RESOLVED, the resolution is a DUPLICATE of bug #219946 https://bugzilla.novell.com/show_bug.cgi?id=219946 The bug is still not fixed. It seems to be a problem with Xserver. Lukas signature.asc Description: OpenPGP digital signature
