Am Freitag, 23. Januar 2004 20:28 schrieb Tom Nielsen:

...

...

My mailserver keeps locking up. Last night at 5:37pm (according to /var/log/messages) was the last time any record of a function was made. Unfortunately the process that was run was not the problem. I can't seem to find out what is causing the problem because it's not listed in /var/log/messages. Is there any where else I might find it?

Last night I tried to get my email, but I couldn't connect. I came in this morning and found my lights on my keyboard blinking and I could log in....so I had to hit the reset button. Once it booted, without any problems might I add, I checked the log and nothing was mentioned.

I'm really stuck on what the problem is. I'm running 8.2.

memtest, badblocks, fsck.

cat, drill, parachute

;-)

Sorry, I didn't get the hints. Are you saying I should run those tests or that they might cause the problems? This always seems to happen at night. I never seem to find file system problems when I reboot.

The reason for your problem can be bad memory, what you can test with memtest. Boot from your CD/DVD and choose memtest at startup. Another reason can be, a broken harddisk or filesystem. First you can test with badblocks, second with fsck. fsck depends on your used filesystem. For reiserfs use reiserfsck. Try both with the rescue-system. I would guess trouble with the filesystem. -- Andreas

Am Freitag, 23. Januar 2004 20:42 schrieb Tom Nielsen:

...

Another reason can be, a broken harddisk or filesystem. First you can test with badblocks, second with fsck. fsck depends on your used filesystem. For reiserfs use reiserfsck. Try both with the rescue-system.

I would guess trouble with the filesystem.

Thanks for the clarification! I really appreciate it. Can I run these tests with the system up and running (as it is now) or is that a bad thing and I should reboot into the rescue system?

I would prefer to boot in the rescue-system for filesystem-checks. -- Andreas

On Fri, 2004-01-23 at 11:47, Patrick Greenwell wrote:

Tom Nielsen wrote:

...

Thanks for the clarification! I really appreciate it. Can I run these tests with the system up and running (as it is now) or is that a bad thing and I should reboot into the rescue system?

Thanks again for the help.

"Very bad things" can happen if you fsck a mounted filesystem. Use the rescue system.

Thanks! That's what I needed to know. I don't like "very bad things". :-) Tom -- Tom Nielsen Neuro Logic Systems 805.389.5435 x18 www.neuro-logic.com

* Tom Nielsen; on 23 Jan, 2004 wrote:

BTW, what's a SuSE guy doing on this? I can't recall the last time I saw someone from SuSE on the list.

there are many but some prefer using their personal mail addresses -- Togan Muftuoglu | Unofficial SuSE FAQ Maintainer | Please reply to the list; http://susefaq.sf.net | Please don't put me in TO/CC. Nisi defectum, haud refiecendum

* Tom Nielsen; on 23 Jan, 2004 wrote:

Ah! That CIA kind of thing.

I do not think it is the paranoia effect, rather preventing the misunderstanding their views represent of SuSE which may not be the case sometimes. -- Togan Muftuoglu | Unofficial SuSE FAQ Maintainer | Please reply to the list; http://susefaq.sf.net | Please don't put me in TO/CC. Nisi defectum, haud refiecendum

On Friday 23 January 2004 2:47 pm, Patrick Greenwell wrote:

...

"Very bad things" can happen if you fsck a mounted filesystem. Use the rescue system.

I've wondered about that. It's obviously true if you make any corrections to the filesystem, and I've seen the warning. But what can go wrong if you fsck a filesystem and leave it undisturbed?

If a filesystem is mounted, it won't have the clean bit on right? Not until it's not live anymore. Ben Yau

...

...

I've wondered about that. It's obviously true if you make any corrections to the filesystem, and I've seen the warning. But what can go wrong if you fsck a filesystem and leave it undisturbed?

If a filesystem is mounted, it won't have the clean bit on right? Not until it's not live anymore.

Seems correct. But then the question is: what's the harm in marking a filesystem as dirty when in fact it's clean?

Interesting. You're asking if fsck checking the filesystem itself intrusive or not? That's an interesting question. We don't do it because there's no point in it really. It will freak out thinking it's not clean and try to fix it or ask us to fix it. In fact I believe some Solaris versions just won't do it and error out since the FS is mounted . But it is an interesting academic question. I vote that you try it and tell us what happens :D I'm guessing that it's still a bad idea. Since it expects the filesystem to be unmounted it may do some intrusive checks. Just a guess.

Well, part of what motivated the question was having an experience of doing just that and later having the filesystem go bad. But that's no proof that the fsck was the cause of the degradation; in fact I doubt it just because I can't see how a purely passive operation would introduce errors. The comment about the clean bit shows that an fsck can have an effect, but even that argument doesn't show how it's a harmful effect. I was running the fsck just to confirm that a filesystem I believed to be OK (I had just created it with a "cp -a") was really OK. Admittedly I was being lazy by not unmounting the filesystem, but I couldn't see why I would need to.

The reason to do it really is any activity on the filesystem will be seen as corruption. As far as it ruining the system, I feel like it would but the more I think about it, you're correct. It would be a passive operation which would incorrectly report corruption on your desk, but wouldn't corrupt the disk itself. Which means this has really become academic since running fsck on a live filesystem wouldn't give you an accurate report anyway. (Althouhg I'm curious. When I get home I'll try it on my RH laptop and see what happens). Ben

Hello. I'm using SuSE 9.0 and LyX 1.3.2, and I can't see math symbols like greek characters or the integral symbol. If I start lyx from an xterm/konsole, the message Adding /usr/share/lyx/xfonts/ to the font path appear in the xterm, but no math symbols are displayed on the lyx window. I tried adding /usr/share/lyx/fonts in a FontPath entry in /etc/X11/XF86Config, then tried use the X font server with that entry on its configuration file, also tried to install the equivalent truetype fonts from latex-xft-fonts, and nothing works. I'm using XFree86 4.3.0. Any ideas?

On Sat, 24 Jan 2004, Togan Muftuoglu wrote:

* Jose Manuel Lara Bauche Alcalde; on 23 Jan, 2004 wrote:

...

Hello. I'm using SuSE 9.0 and LyX 1.3.2, and I can't see math symbols like greek characters or the integral symbol. If I start lyx from an xterm/konsole, the message Adding /usr/share/lyx/xfonts/ to the font path appear in the xterm, but no math symbols are displayed on the lyx window. I tried adding /usr/share/lyx/fonts in a FontPath entry in /etc/X11/XF86Config, then tried use the X font server with that entry on its configuration file, also tried to install the equivalent truetype fonts from latex-xft-fonts, and nothing works. I'm using XFree86 4.3.0.

have you tried running "lyx -dbg 512" for other options try lyx -dbg

I ran lyx -dbg 512 and got this: Setting debug level to 512 Debugging `font' (Font handling) Font 'Sans Serif, Bold, Upright, Normal, yellow, Emphasis Off, Underline Off, Noun Off, Language: Spanish' matched by xft The font has size: 15 This font is an exact match XFLD: xft Font 'Roman, Medium, Upright, Normal, none, Emphasis Off, Underline Off, Noun Off, Language: Spanish' matched by xft The font has size: 15 This font is an exact match XFLD: xft Looking for font family eufm10 ... FAILED :-( Looking for font family eufm10 ... FAILED :-( Looking for font family cmsy10 ... FAILED :-( Looking for font family cmmi10 ... FAILED :-( Looking for font family symbol ... FAILED :-( Looking for font family cmr10 ... FAILED :-( Looking for font family cmex10 ... FAILED :-( Looking for font family msam10 ... FAILED :-( Looking for font family msbm10 ... FAILED :-( Looking for font family wasy10 ... FAILED :-(

On Sat, 24 Jan 2004, Togan Muftuoglu wrote:

* Jose Manuel Lara Bauche Alcalde; on 23 Jan, 2004 wrote:

...

On Sat, 24 Jan 2004, Togan Muftuoglu wrote:

Looking for font family eufm10 ... FAILED :-(

pin eufm10

/CD1/suse/i586/tetex-2.0.1-22.i586.rpm: -rw-r--r-- root root 1787 Jan 4 1995 /usr/share/texmf/fonts/source/ams /euler/eufm10.mf

so looks lke you have the tetex package missing (probably other tex related packages as well

tetex is not missing: pin eufm10 ./CD1/suse/i586/cjk-lyx-1.3.2-72.i586.rpm: lrwxrwxrwx root root 53 Nov 5 02:45 /usr/share/lyx-cjk/xfonts/euf m10.pfb -> /usr/share/texmf/fonts/type1/bluesky/euler/eufm10.pfb ./CD1/suse/i586/tetex-2.0.2-76.i586.rpm: -rw-r--r-- root root 1787 Jan 4 1995 /usr/share/texmf/fonts/source/a ms/euler/eufm10.mf ./CD1/suse/i586/tetex-2.0.2-76.i586.rpm: -rw-r--r-- root root 1040 Aug 14 1995 /usr/share/texmf/fonts/tfm/ams/ euler/eufm10.tfm ./CD1/suse/i586/tetex-2.0.2-76.i586.rpm: -rw-r--r-- root root 25453 Jun 18 1997 /usr/share/texmf/fonts/type1/bl uesky/euler/eufm10.pfb ./CD1/suse/i586/mupad-2.5.2-7.i586.rpm: -rw-r--r-- root root 5048 Nov 5 2002 /opt/MuPAD/share/doc/dvi-de/font s/eufm10.329pk ./CD1/suse/i586/mupad-2.5.2-7.i586.rpm: -rw-r--r-- root root 5048 Nov 5 2002 /opt/MuPAD/share/doc/dvi/fonts/e ufm10.329pk

On Sat, 24 Jan 2004, Togan Muftuoglu wrote:

* Jose Manuel Lara Bauche Alcalde; on 23 Jan, 2004 wrote:

...

tetex is not missing:

I beg to differ pin is showing you the location of tetex in the CD

If you had it installed this would be what pin displaying

pin 0.27 - package info for tetex

------------------------------------------------------------------ *** rpm info ------------------------------------------------------------------

Name : tetex Relocations: (not relocateable) Version : 2.0.1 Vendor: SuSE Linux AG, Nuernberg, Germany Release : 22 Build Date: Mon Mar 17 15:03:53

you can verify it (that it is not installed rpm -q tetex

It is installed: rpm -q tetex tetex-2.0.2-76 rpm -ql tetex | grep eufm10 /usr/share/texmf/fonts/source/ams/euler/eufm10.mf /usr/share/texmf/fonts/tfm/ams/euler/eufm10.tfm /usr/share/texmf/fonts/type1/bluesky/euler/eufm10.pfb ls /usr/share/texmf/fonts/type1/bluesky/euler euex10.pfb eufb5.pfb eufm10.pfb eufm7.pfb eurb5.pfb eurm10.pfb eurm7.pfb eusb5.pfb eusm10.pfb eusm7.pfb eufb10.pfb eufb7.pfb eufm5.pfb eurb10.pfb eurb7.pfb eurm5.pfb eusb10.pfb eusb7.pfb eusm5.pfb

On Fri, 23 Jan 2004, Jose Manuel Lara Bauche Alcalde wrote:

On Sat, 24 Jan 2004, Togan Muftuoglu wrote:

...

* Jose Manuel Lara Bauche Alcalde; on 23 Jan, 2004 wrote:

...

Hello. I'm using SuSE 9.0 and LyX 1.3.2, and I can't see math symbols like greek characters or the integral symbol. If I start lyx from an xterm/konsole, the message Adding /usr/share/lyx/xfonts/ to the font path appear in the xterm, but no math symbols are displayed on the lyx window. I tried adding /usr/share/lyx/fonts in a FontPath entry in /etc/X11/XF86Config, then tried use the X font server with that entry on its configuration file, also tried to install the equivalent truetype fonts from latex-xft-fonts, and nothing works. I'm using XFree86 4.3.0.

have you tried running "lyx -dbg 512" for other options try lyx -dbg

I ran lyx -dbg 512 and got this:

Setting debug level to 512 Debugging `font' (Font handling) Font 'Sans Serif, Bold, Upright, Normal, yellow, Emphasis Off, Underline Off, Noun Off, Language: Spanish' matched by xft The font has size: 15 This font is an exact match XFLD: xft Font 'Roman, Medium, Upright, Normal, none, Emphasis Off, Underline Off, Noun Off, Language: Spanish' matched by xft The font has size: 15 This font is an exact match XFLD: xft Looking for font family eufm10 ... FAILED :-( Looking for font family eufm10 ... FAILED :-( Looking for font family cmsy10 ... FAILED :-( Looking for font family cmmi10 ... FAILED :-( Looking for font family symbol ... FAILED :-( Looking for font family cmr10 ... FAILED :-( Looking for font family cmex10 ... FAILED :-( Looking for font family msam10 ... FAILED :-( Looking for font family msbm10 ... FAILED :-( Looking for font family wasy10 ... FAILED :-(

By the way... my /usr/share/lyx/xfonts directory has this contents: cmex10.pfb -> /usr/share/texmf/fonts/type1/bluesky/cm/cmex10.pfb cmmi10.pfb -> /usr/share/texmf/fonts/type1/bluesky/cm/cmmi10.pfb cmr10.pfb -> /usr/share/texmf/fonts/type1/bluesky/cm/cmr10.pfb cmsy10.pfb -> /usr/share/texmf/fonts/type1/bluesky/cm/cmsy10.pfb encodings.dir eufm10.pfb -> /usr/share/texmf/fonts/type1/bluesky/euler/eufm10.pfb fonts.dir fonts.scale msam10.pfb -> /usr/share/texmf/fonts/type1/bluesky/symbols/msam10.pfb msbm10.pfb -> /usr/share/texmf/fonts/type1/bluesky/symbols/msbm10.pfb wasy10.pfb -> /usr/share/texmf/fonts/type1/hoekwater/wasy/wasy10.pfb and symlinks are not broken.

On Saturday 24 January 2004 00:20, Jose Manuel Lara Bauche Alcalde wrote:

I updated 8.2 to 9.0 and now when I run YOU, system information says:

Product: SuSE Linux Version: 8.2 Base Architecture: i386

I'm afrid YOU will try to get 8.2 patches instead of 9.0.

If let YOU check for new updates and accept them, I get Error: Signature check failed for every patch it tries to download.

How can I correct this?

Edit /var/adm/YaST/ProdDB/prod_0000001 and change Product data to SuSE Linux-9.0- SuSE-Linux-Professional-i386-9.0-0 and YouPath to i386/update/9.0 I *think* (AFAICR at least) that's all you need to change

On Saturday 24 January 2004 00:56, Jose Manuel Lara Bauche Alcalde wrote:

Well... I sent you a message to your personal mail instead to the list... It was a mistake... And the message was wrong... I thought everything was right, but no. Version is right, now, and YOU downloads the patches... But when it tries to apply the patch, it still returns Error: Signature check failed.

Make sure the package suse-build-key is installed

The Friday 2004-01-23 at 11:42 -0800, Tom Nielsen wrote:

...

I would guess trouble with the filesystem.

Thanks for the clarification! I really appreciate it. Can I run these tests with the system up and running (as it is now) or is that a bad thing and I should reboot into the rescue system?

Surprisingly, some of them can. If there is a hardware problem with your harddisk, badblocks, sector relocation, etc (not filesystem), and your HD and BIOS support SMART (Self-Monitoring, Analysis and Reporting Technology), then: a) You can see the log of previous hard disk errors. b) You can initiate testing the HD, while linux is running normally. b) you can view the tests result later (+/-60'). How? With "smartctl" (included on your SuSE CDs/DVD). Launch short test: smartctl --test=short /dev/hda Launch long test (not simultaneously to short test!): smartctl --test=long /dev/hda View test log: smartctl --log=selftest /dev/hda View all info - includin error log: smartctl --all /dev/hda View health status of disk: smartctl --health /dev/hda More info: man smartctl -- Cheers, Carlos Robinson

On Fri, 2004-01-23 at 15:12, Anders Johansson wrote:

On Friday 23 January 2004 20:35, Andreas Winkelmann wrote:

...

Another reason can be, a broken harddisk or filesystem. First you can test with badblocks, second with fsck. fsck depends on your used filesystem. For reiserfs use reiserfsck. Try both with the rescue-system.

Not all hard disk problems can be detected with tools though. One system I used to have almost always crashed overnight, and there was never any visible problem with the hard drive. Until, that is, it finally gave up completely and refused to even power up. Before that happened, every tool said everything was just peachy. Once a new hdd was in, the system was stable as a rock

Oh, ya. That's encouraging. :-) The strange thing is I have a mirrored raid on this machine, so shouldn't it keep right on trucking if something happens to the other drive? Tom -- Tom Nielsen Neuro Logic Systems 805.389.5435 x18 www.neuro-logic.com

WOW! That's a good idea! Thanks! Tom Sorry about all the ! but I mean it. On Fri, 2004-01-23 at 12:53, John Andersen wrote:

On Friday 23 January 2004 10:21, Tom Nielsen wrote:

...

Last night I tried to get my email, but I couldn't connect. I came in this morning and found my lights on my keyboard blinking and I could log in....so I had to hit the reset button. Once it booted, without any problems might I add, I checked the log and nothing was mentioned.

I'm really stuck on what the problem is. I'm running 8.2.

Thanks for any ideas! Tom

Tom... Blinking lights are kernel panic... Switch to console 10 (ctrl-alt-f10) and leave it there, and that way you will at least get to see the last screen full of messages after a panic.

-- _____________________________________ John Andersen

-- Tom Nielsen Neuro Logic Systems 805.389.5435 x18 www.neuro-logic.com