On Tuesday 16 January 2007 14:03, Patrick Shanahan wrote:

...

remember Goggle ???

As in "beer goggles?"

... Patrick Shanahan

RRS -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

* Randall R Schulz [01-16-07 17:20]:

On Tuesday 16 January 2007 14:03, Patrick Shanahan wrote:

...

remember Goggle ???

As in "beer goggles?"

I believe that I could get much cheaper eye-glasses if I just used to Budweiser glass elements :^) Pairs makes the daz go bettter :^) -- Patrick Shanahan Registered Linux User #207535 http://wahoo.no-ip.org @ http://counter.li.org HOG # US1244711 Photo Album: http://wahoo.no-ip.org/gallery2 OpenSUSE Linux http://en.opensuse.org/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

ON Tuesday, January 16, 2007 @ 4:04 PM, Patrick Shanahan wrote:

* Greg Wallace [01-16-07 15:17]:

...

I have been running 10.2 for several weeks now and just in the last two or three days my system ran fsck at each startup. I've only rebooted once a day, so I have no idea why this started happening. I don't believe it was happening until just recently, so maybe it's a result of a system patch (?). I forget where you set the count as to how many reboots need to occur between fscks. Could someone tell me where that setting is? I know I haven't modified it and for as long as I have had SuSE (since 8.1) I believe it was set at 99. It's never done this before.

remember Goggle ???

a goggle search for "file system check frequency" eighth suggestion gives "tune2fs" which exists in e2fsprogs, for ext2/3 filesystems (you didn't say which). There is a manpage.

I'm sure that a similar utility exists for other filesystems. Goggle is available.

I had tried Google, but couldn't come up with a phrase that narrowed things down enough. Mostly got hits about how to run fsck. I tried your search criteria and found the one about tune2fs. I ran -- tune2fs -T and got ...(29-May-2006). That's the same date I see in the log every time fsck runs. Sounds like a problem. I'm going to look at this command some more and see if I can dig any other information out. However, it would seem that running e2fsck would automatically re-set that date, but maybe there is some special type of e2fsck running on my machine that isn't driven by date. On the other hand, 29-May-2006 is a long time ago. Surely an auto-fsck should have been run since then. Pretty strange. Greg W -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Tuesday 2007-01-16 at 17:49 -0600, Greg Wallace wrote:

...

else - and it may tell what.

...

Perhaps an fstab incosistency, a "/forcefsck" file...

That's right, of course. I'd completely forgotten that that message came out when it was fscking on cycle. I have not gotten that message. One thing I have noticed each time is a message that says --

fsck 1.39 (29-May-2006)

I'd guess that's the version string ;-) nimrodel:~ # fsck --version fsck 1.38 (30-Jun-2005) fsck.ext3: invalid option -- e Usage: fsck.ext3 [-panyrcdfvstDFSV] [-b superblock] [-B blocksize] ...

That still wouldn't explain why the message "filesytem not checked..." isn't appearing. Maybe that message doesn't appear any longer under 10.2?

My guess is that it doesn't appear because it doesn't have to. I mean, fsck is not running for that reason, but for another one. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFFrXcutTMYHG2NR9URAjHwAJ9rA20+dQSZoLVjW53bt6fyaRrLVACfZvTU zN7xlaF3z5cuxSZhyCSQW10= =LXAh -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Greg Wallace wrote:

On Tuesday, January 16, 2007 @ 7:09 PM, Carlos Robinson wrote:

...

The Tuesday 2007-01-16 at 17:49 -0600, Greg Wallace wrote:

...

...

...

else - and it may tell what.

Perhaps an fstab incosistency, a "/forcefsck" file...

That's right, of course. I'd completely forgotten that that message came out when it was fscking on cycle. I have not gotten that message. One thing I have noticed each time is a message that says --

fsck 1.39 (29-May-2006)

...

I'd guess that's the version string ;-)

...

nimrodel:~ # fsck --version fsck 1.38 (30-Jun-2005)

You're right, that is the version.

...

fsck.ext3: invalid option -- e Usage: fsck.ext3 [-panyrcdfvstDFSV] [-b superblock] [-B blocksize] ...

...

...

That still wouldn't explain why the message "filesytem not checked..." isn't appearing. Maybe that message doesn't appear any longer under 10.2?

...

My guess is that it doesn't appear because it doesn't have to. I mean, fsck is not running for that reason, but for another one.

There's definitely a problem here. It is indeed running every time I boot. Not only that, but I went to look at the messages file and it's so large that YAST can't even bring it up (it just quits). I took a look at it and it has 1,128,168,967 for a size under ls. I'm trying to remember how to use less and have it start at the end of the file and go backward so I can see what some of the last messages in the file are.

...

- -- Cheers, Carlos E. R.

Greg Wallace

Why don't you use tail instead. To find more features just type "man tail" -- Joseph Loo jloo@acm.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Tuesday 2007-01-16 at 20:58 -0600, Greg Wallace wrote:

I want to be able to scroll backward from the end of the file, not just see the last few lines in the file. I know there's a way to do it because I've done it before. I just have to remember how. I thought there was a command under less that said go to the end, but I can't find any reference to it under "man less".

Just type "h" for help under less. The key [End] goes to the end of the file. As the file is large, it will take sometime to load first (calculating line numbers) before it allows you to browse. Or (man rules!): + If a command line option begins with +, the remainder of that option is taken to be an initial command to less. For example, +G tells less to start at the end of the file rather than the beginning, and +/xyz tells it to start Or you might want to use "dog" instead of "cat" ;-) - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFFrZQvtTMYHG2NR9URAoYQAJ0QIrfI3p1H5pDIRD2296BVvBEkcACfe7GT 5vdvwAJhWC1WvWGydwbNj7Y= =ZYoH -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Tuesday, January 16, 2007 @ 10:29 PM, Wade Jones wrote:

Tuesday 2007-01-16 at 20:58 -0600, Greg Wallace wrote:

...

I want to be able to scroll backward from the end of the file, not just see the last few lines in the file. <snip>

On Tuesday 16 January 2007 21:12, Carlos E. R. wrote: Or you might want to use "dog" instead of "cat" ;-)

-- Cheers, Carlos E. R.

=============================================

Or you could use tac..................................................

NAME tac - concatenate and print files in reverse

Wade

Wow! What a super command. Takes you to the end of the file instantly! Easy to remember to (cat spelled backwards). Thanks for the tip! Greg Wallace -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Tuesday 16 January 2007 18:58, Greg Wallace wrote:

On Tuesday, January 16, 2007 @8:53 PM, Joseph Loo wrote:

<snip>

...

Why don't you use tail instead. To find more features just type "man tail"

--

Joseph Loo jloo@acm.org

I want to be able to scroll backward from the end of the file, not just see the last few lines in the file. I know there's a way to do it because I've done it before. I just have to remember how. I thought there was a command under less that said go to the end, but I can't find any reference to it under "man less".

Greg W.

You press the "b" key and it will have less or more go back one page. Mike -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Tuesday, January 16, 2007 @ 11:09 PM, Randall Schulz wrote:

On Tuesday 16 January 2007 18:58, Greg Wallace wrote:

...

On Tuesday, January 16, 2007 @8:53 PM, Joseph Loo wrote:

<snip>

...

Why don't you use tail instead. To find more features just type "man tail"

--

Joseph Loo jloo@acm.org

I want to be able to scroll backward from the end of the file, not just see the last few lines in the file. I know there's a way to do it because I've done it before. I just have to remember how. I thought there was a command under less that said go to the end, but I can't find any reference to it under "man less".

Less's commands are vaguely similar to Vi's and / or Vim's. To go to the last line type 'G' (capital 'g'). To go to the first line, use lower-case 'g'.

You can get help while less is running by typing 'h'.

...

Greg W.

Randall Schulz

Thanks for the tips. Greg W -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On 2007-01-17 15:24, Greg Wallace wrote:

Well, I've done some more digging on this problem and, though I have found out some things, I am still no closer to solving the problem than before.

1) The huge amount of lines in /var/log/messages seems to be unrelated to why I'm getting an fsck every time I boot. There are hundreds of lines in there that look like this one --

Linux kernel: SFW2-IN-ACC-RELATED IN eth0 OUT=MAC=00:08:74:24:85:82:00:04:5A:0f:18:07:08:00 SRC=128.61.111.11 DST=192.168.1.102 LEN=529 TOS=0x00 PREC=0x00 TTL=64 ID=238936 WINDOW=1716 RES=0x00 ACK FIN URGP=0 OPT (long hex number here)

This is firewall logging. Why it is in /var/log/messages is a mystery, because (at least in 9.3) the syslog-ng config file is written to write all firewall log entries to a separate file. Possibly you are still using the old syslog, which is pretty brain dead and dumps everything from the kernel into /var/log/messages. If so, I suggest you dump it and install syslog-ng instead. -- The best way to accelerate a computer running Windows is at 9.81 m/s² -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On 2007-01-17 17:40, Greg Wallace wrote:

On Wednesday, January 17, 2007 @ 4:30 PM, Darryl Gregorash wrote:

...

On 2007-01-17 15:24, Greg Wallace wrote:

<snip>

...

Linux kernel: SFW2-IN-ACC-RELATED IN eth0 ^^^

...

This is firewall logging. Why it is in /var/log/messages is a mystery, I may have found it. Try this:

grep SFW2 /var/log/messages | grep IN= Then this: grep IN= /etc/syslog-ng/syslog-ng.conf.in The second command will probably output a line like this: filter f_iptables { facility(kern) and match("IN=") and match("OUT="); }; However, in the log entry you posted, the text is "IN", not "IN=". No match, so subsequent rules dump the entry to /var/log/messages. This is possibly a bug in the iptables logging module, ipt_LOG, for that kernel version. BTW, what is the result of this: ls -l /var/log/firewall If it is zero size, or doesn't exist at all, you can just edit /etc/syslog-ng/syslog-ng.conf.in to read "IN " vs. "IN=", run 'SuSEconfig --module syslog-ng', and carry on (until the hiccup is fixed, then you'd have to reverse the change :-) ). Note: edit the .conf.in file, not the .conf file, or you will lose the changes when suseconfig is run. -- The best way to accelerate a computer running Windows is at 9.81 m/s² -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On 2007-01-17 22:10, Greg Wallace wrote:

On Wednesday, January 17, 2007 @ 8:24 PM, Darryl Gregorash wrote:

<snip>

...

grep SFW2 /var/log/messages | grep IN=

Thousands of lines of output

OK, so much for that idea. I'm sure glad, though, that no bug report has to be filed against the kernel :-)

...

grep IN= /etc/syslog-ng/syslog-ng.conf.in

No such file or directory

Say what? What's the output of rpm -ql syslog-ng

...

BTW, what is the result of this: ls -l /var/log/firewall

ls: cannot access /var/log/firewall: No such file or directory

Maybe that's the problem. There is no separate log file set up for the firewall so all of the firewall messages get dumped into messages. Is there somewhere where you can define a specific separate log file for firewall messages? If the file does not exist, the syslog daemon will create it -- assuming, of course, that it is given any reason to do so.

-- The best way to accelerate a computer running Windows is at 9.81 m/s² -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On 2007-01-18 01:02, Greg Wallace wrote:

On Wednesday, January 17, 2007 @ 11:51 PM, Darryl Gregorash wrote:

...

<snip>

...

...

grep IN= /etc/syslog-ng/syslog-ng.conf.in

No such file or directory

Say what? What's the output of

rpm -ql syslog-ng

A list of about 40 files

I'm interested only in what is *not* in /usr/share/doc/ and /usr/share/man/. That should be only about 4 files or so. In particular, is there a syslog-ng.conf.in anywhere? Or a syslog-ng.conf?

So having thousands of lines of output generated by the firewall isn't reason enough to create the file? How can I give it a reason to create it?

No amount of output is sufficient to create a file, if creating the file is not specified in the syslog-ng config file. -- The best way to accelerate a computer running Windows is at 9.81 m/s² -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On 2007-01-18 12:07, Greg Wallace wrote:

On Thursday, January 18, 2007 @ 1:50 AM, Darryl Gregorash wrote:

...

<snip> is there a syslog-ng.conf.in anywhere? Or a syslog-ng.conf?

Ok. I see. Yes, there are two files meeting the above criteria --

/etc/syslog-ng/syslog-ng.conf /etc/syslog-ng/syslog-ng.conf.rpmnew

OK, so much for assuming the config files in your version are the same as in mine. Apologies for the confusion. Please check in the Yast sysconfig editor for the values of the following (under System/Logging): SYSLOG_DAEMON SYSLOG_NG_CREATE_CONFIG The first should be set to "syslog-ng", but it may be set to "syslogd" (that would explain right away why there is no /var/log/firewall). The second might not even be present in your system, but if it does exist, I need to know its value to be able to tell you which file (if any) you need to edit -- we don't want any changes disappearing the next time SuSEconfig runs :-) -- The best way to accelerate a computer running Windows is at 9.81 m/s² -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On 2007-01-18 14:05, Greg Wallace wrote:

On Thursday, January 18,, 2007 @ 1:33 PM, Darryl Gregorash wrote:

<snip>

...

The first should be set to "syslog-ng", but it may be set to "syslogd" (that would explain right away why there is no /var/log/firewall).

Bingo! It was set to syslog. I clicked on the drop-down and changed it to syslog-ng

<snip> I don't have SYSLOG_NG_CREATE_CONFIG. Hopefully, changing the value of SYSLOG_DAEMON will do the trick.

Well, it will certainly get you running with syslog-ng :-) -- but perhaps it is configured to put all firewall messages into /var/log/messages anyway :-) If you now have a /var/log/firewall file, this one is done. Alas, it doesn't answer any of your questions about fsck, but now your mail kernel log isn't going to be full of firewall traffic :-) -- The best way to accelerate a computer running Windows is at 9.81 m/s² -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On 2007-01-18 16:42, Greg Wallace wrote:

<snip>

I do indeed now have a /var/log/firewall file. I took a look at its contents and it looks just like what was going into messages. So, maybe that is fixed. Now what I'd like to do is rotate out the current huge messages file and start with a new one. Can you tell me how to safely do that? As root, run "logrotate -f /etc/logrotate.conf"

That will rotate all logs, even those that do not qualify as per the logrotate configuration, create new (empty) files for each, and restart any service that needs to be restarted. Logging will then continue normally. If you only wish to rotate /var/log/messages but nothing else, create a file containing the following: /var/log/messages { compress dateext maxage 365 rotate 99 missingok notifempty size +4096k create 640 root root sharedscripts postrotate /etc/init.d/syslog reload endscript } (I actually just extracted this from /etc/logrotate.d/syslog, and removed all the log file names except /var/log/messages.) You can name this file anything you wish, filename is as good as any, and place it anywhere you want, such as /tmp :-) Then run "logrotate -f /tmp/filename", again as root. -- The best way to accelerate a computer running Windows is at 9.81 m/s² -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Greg Wallace wrote:

I do indeed now have a /var/log/firewall file. I took a look at its contents and it looks just like what was going into messages. So, maybe that is fixed. Now what I'd like to do is rotate out the current huge messages file and start with a new one. Can you tell me how to safely do that?

Thanks, Greg Wallace

Something you might like to consider, now you're using syslog-ng, is getting it to have a different log file for each day. Here are the "destination" and "log" entries in my syslog-ng config file: ############################################################### destination syslog {file("/var/log/$FACILITY.log.$YEAR$MONTH$DAY");}; destination full-syslog {file("/var/log/system.log.$YEAR$MONTH$DAY"); ############################################################### log {source(src); destination(syslog); }; log {source(src); destination(full-syslog); }; ############################################################### What this will give you is, for example, /var/log/system.log.20070119 which will contain all the syslog messages, and files like /var/log/local0.log.20070119, /var/log/mail.log.20070119 etc. which will contain the messages for each facility. It makes tracking things down a bit easier. Rather than use logrotate (most of my systems are Solaris) I use the following shell script to gzip the log files and put them in /var/log/archive (Which I have on a separate filesystem). It's run at 23:59 each night: ====cut here=== #!/bin/bash cd /var/log # Get TODAY's date. As we're running just before midnight, go to # sleep until after midnight so that the log files will no longer # be being written to DAT=`date "+%Y%m%d"` sleep 65 # As we're running from cron, output a friendly message to say what we're doing echo "Compressing and archiving log files: \n" # get a single colum list of all the log files that were generated then gzip # each file before moving to /var/logs/archive ls -1 *.log.${DAT} | while read fil do echo $fil gzip $fil mv ${fil}.gz /var/log/archive done # Log files are retained for a maximum of 1 year/366 days (to cater for leap years) # We remove everything older than 345 days because our oldest backups are 21 days old # and will have 345 days' worth of logs on them echo "Removing archived logs older than 345 days:\n " cd /var/log/archive # the above cd isn't really necessary as the find command explicitly states where to search. # DON'T be tempted to do find . -name "*.log.gz" ... you might just regret it! :) # Find all the log files older than 345 days and remove them. Echo the file name so it gets # captured in the cron output find /var/log/archive -name "*.log.*.gz" -mtime +345 | while read arc do rm ${arc} echo ${arc} done ====cut here=== The crontab entry is: 59 23 * * * /usr/local/scripts/clear_logs (I keep my home-brewed stuff in /usr/local/scripts) Hope that helps - -- Paul Walsh -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org iD8DBQFFsJ7EMgcRY6KQRjQRAtVJAJ9mnZwwCAHBtRz1IX5rv6BZEfzYCACfQncE POYiy5wvkSkiUNMLUPucTIU= =8GlG -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Wednesday 2007-01-17 at 22:10 -0600, Greg Wallace wrote:

...

Then this:

...

grep IN= /etc/syslog-ng/syslog-ng.conf.in

No such file or directory

That is your problem. Your syslog-ng installation is broken, repair it. Then the messages from the firewall will be sorted to the appropriate file. Try this: grep -i SYSLOG_DAEMON /etc/sysconfig/syslog you should get: SYSLOG_DAEMON="syslog-ng" If you get "syslog", then fire up the Yast package module, uninstall syslog, and install syslog-ng instead. If you get "syslog-ng", I would reinstall it.

ls: cannot access /var/log/firewall: No such file or directory

Maybe that's the problem. There is no separate log file set up for the firewall so all of the firewall messages get dumped into messages. Is there somewhere where you can define a specific separate log file for firewall messages?

See above. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFFr0vktTMYHG2NR9URAkaRAJwOUUNkDXp3i2YMJyUtQoFRQz6unACfYZD6 aTpM4NXtK1RR5/p51ZBqxZA= =aHcc -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

* Greg Wallace [01-18-07 13:39]:

On Thursday, January 18, 2007 @ 4:29 AM, Carlos Robinson wrote: [..]

...

If you get "syslog", then fire up the Yast package module, uninstall syslog, and install syslog-ng instead.

...

If you get "syslog-ng", I would reinstall it.

I tried re-installing syslog-ng and still get "syslogd" as output from the above command. By the way, I have both syslogd and syslog-ng installed. Is that what I should have?

Gregg, you are not reading the instructions. Carlos said: If you get "syslog", then fire up the Yast package module, uninstall syslog, and install syslog-ng instead. and If you get "syslog-ng", I would reinstall it. You REALLY need to read the instructions. -- Patrick Shanahan Registered Linux User #207535 http://wahoo.no-ip.org @ http://counter.li.org HOG # US1244711 Photo Album: http://wahoo.no-ip.org/gallery2 OpenSUSE Linux http://en.opensuse.org/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

* Darryl Gregorash [01-18-07 14:44]:

Hey, don't blame the patient :-)

If anyone, blame me -- I haven't actually been too wide awake on this one.

the instructions were not vague. He has two versions installed, syslog and syslog-ng. Contributing to confusion in his chair. He is grasping straws and not comprehending instructions as written. Perhaps, the rest should just keep quiet. You take him thru the steps. You are very capable. :^) -- Patrick Shanahan Registered Linux User #207535 http://wahoo.no-ip.org @ http://counter.li.org HOG # US1244711 Photo Album: http://wahoo.no-ip.org/gallery2 OpenSUSE Linux http://en.opensuse.org/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

* Greg Wallace [01-18-07 15:01]:

I got syslogd, not syslog, but I re-installed syslog-ng anyway, as I said above. I think that pretty much matches what I was instructed to do (I didn't get syslog so I re-installed syslog-ng). If not, what did I miss?

I believe that you have two versions, syslog and syslog-ng. IIANM, syslogd belongs to syslog. Do "rpm -qf `which syslogd` and you will see. You are receiving toooo many instructions and that has to be confusing you. Darryl Gregorash is quite knowledgable, more so than I. Let him take you thru the steps. I am gone from this thread and all related. gud luk, -- Patrick Shanahan Registered Linux User #207535 http://wahoo.no-ip.org @ http://counter.li.org HOG # US1244711 Photo Album: http://wahoo.no-ip.org/gallery2 OpenSUSE Linux http://en.opensuse.org/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Thursday, January 18, 2007 @ 6:19 PM, Carlos Robinson wrote:

The Thursday 2007-01-18 at 12:35 -0600, Greg Wallace wrote:

...

...

Try this:

...

grep -i SYSLOG_DAEMON /etc/sysconfig/syslog

...

you should get:

...

SYSLOG_DAEMON="syslog-ng"

I get SYSLOG_DAEMON="syslogd".

...

If you get "syslog", then fire up the Yast package module, uninstall syslog, and install syslog-ng instead.

...

If you get "syslog-ng", I would reinstall it.

I tried re-installing syslog-ng and still get "syslogd" as output from the above command. By the way, I have both syslogd and syslog-ng installed. Is that what I should have?

I said above "uninstall syslog" (or syslogd, same thing). Do uninstall it, please. Keep only syslog-ng. Then make sure you have

SYSLOG_DAEMON="syslog-ng"

- -- Cheers, Carlos E. R.

Darryl Gregorash already had me switch the SYSLOG_DAEMON to syslog-ng and that seemed to fix the problem. However, I hadn't uninstalled syslogd. I just did that. Just switching the daemon to syslog-ng fixed the messages file problem, but if syslog-ng supplants syslogd then it will now be out of the way so as to not possibly cause problems down the road. Thanks, Greg W -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Thursday 2007-01-18 at 22:42 -0600, Darryl Gregorash wrote:

...

<snip> ... SYSLOG_DAEMON to syslog-ng and that seemed to fix the problem. This alone was sufficient; there are no conflicting files in the two

On 2007-01-18 20:56, Greg Wallace wrote: packages. Deleting syslogd avoids the possibility that it will somehow be selected again sometime in the future, but that is all.

Right. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFFsKXmtTMYHG2NR9URAk8OAJ9kwICu7kd7M7JxodLJjkSrrmsYXQCcD0+/ DJMmR4NZ4JdVVgihF7iY0Ow= =kpl/ -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Friday 2007-01-19 at 14:44 -0600, Greg Wallace wrote:

...

...

...

... SYSLOG_DAEMON to syslog-ng and that seemed to fix the problem. This alone was sufficient; there are no conflicting files in the two packages. Deleting syslogd avoids the possibility that it will somehow be selected again sometime in the future, but that is all.

That's pretty much what I figured, but no reason to have a package installed that isn't of any use. That being the case, it seemed to be a good idea to just go ahead and trash it.

Absolutely. And you save some disk space, and some update resources if Yast want's to update something you don't use, and less confussions... Yes, the right thing is removing it. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFFsWlctTMYHG2NR9URAoYHAJ42wsCrPXpPcczR15zKgTwbKhUi4gCeIkvH wla6GDTQZxl4QKnmG0AW0s0= =t2s3 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Friday, January 19, 2007 @ 8:35 PM, Darryl Gregorash wrote:

On 2007-01-19 18:59, Carlos E. R. wrote:

...

<snip>

And you save some disk space, and some update resources if Yast want's to update something you don't use, and less confussions... Yes, the right thing is removing it.

And from now on, knowing what he does, the correct thing is not to let it install in the first place -- during package selection when installing/upgrading.

You probably heard a rather loud noise coming from this direction when syslog-ng became available as an alternative to syslogd. That was just me cheering wildly :-)

-- This space for rent. Inquire within.

Well, I have always done upgrade installs because I have certain software that is spread throughout the directories of the system (Oracle, for instance) that would be very difficult to re-install from scratch (difficult for me anyway). So, the upgrade process says something like there are n number of packages being deleted. True you can go in and look at the details and see everything being deleted, added, etc., but how, for instance, would the average SUSE "user" know that syslog-ng was a replacement for syslogd. As a matter of fact, since it was a replacement, the installation process should have trashed it automatically, like it does for most packages that are becoming obsolete for that upgrade. For those doing clean installs this is a non-issue, but when upgrading you're pretty much relying on that upgrade process to handle these things. Understandably, constructing an upgrade procedure that can correctly handle all combinations and delete just what needs deleting in all cases, etc. would be pretty much impossible, so you're left with some situations like this one, where an obsolete package was left in place along with the new package that was supposed to take its place. Just something you have to live with if you're not doing clean installs I guess. Greg Wallace -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On 2007-01-19 21:45, Greg Wallace wrote:

...

<snip> details and see everything being deleted, added, etc., but how, for instance, would the average SUSE "user" know that syslog-ng was a replacement for syslogd. As a matter of fact, since it was a replacement, the installation process should have trashed it automatically, like it does for most packages that are becoming obsolete for that upgrade. For those doing clean installs this is a non-issue, but when upgrading you're

On Friday, January 19, 2007 @ 11:07 PM, Darryl Gregorash wrote: pretty

...

much relying on that upgrade process to handle these things.

Syslogd is probably still there because many use it for whatever reason. If you have it installed, it is up to you, during the upgrade, to specify it should be replaced by any alternative. This is the same as any installed package for which an alternative exists. It is not possible for the distribution authors to know which package(s) should or should not be replaced, nor is it acceptable IMO for them to attempt to presume any such knowledge. I most certainly would scream very loudly if, during a full system upgrade, the installation process decided that no, I do not actually want to use Seamonkey, rather I wish instead to use Firefox and Thunderbird (or Konqueror and Evolution, or .....). I am quite certain that you would complain even more loudly if the installation decided that you didn't need Oracle, but wanted MySQL instead.

True enough. Since syslog-ng is basically a new improved version of syslogd it seemed that it would be only logical for the system to make that choice. But, as you say, some might not want to and it would be presumptuous for the system to make that assumption. And both can co-exist without any harm being done, allowing one to even switch back and forth between the two, if anyone would really want to do that. So, I stand corrected. Greg W -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Wed, 2007-01-17 at 15:24 -0600, Greg Wallace wrote:

Well, I've done some more digging on this problem and, though I have found out some things, I am still no closer to solving the problem than before.

1) The huge amount of lines in /var/log/messages seems to be unrelated to why I'm getting an fsck every time I boot. There are hundreds of lines in there that look like this one --

Just a thought here...wouldn't the messages of interest be in /var/log/boot.msg instead of /var/log/messages? That's where all my reiserfs messages are, before the kernel is booted up... Tom in NM -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On 2007-01-17 16:46, Greg Wallace wrote:

<snip> messages to find any pertinent info. I found some lines ahead of where the fsck is invoked relating to resume from disk. Seems like that process is failing. I don't even need resume from disk, as far as I know, so I'm

In my /boot/grub/menu.lst: kernel (hd0,0)/vmlinuz root=/dev/hda2 vga=normal selinux=0 splash=silent resume=/dev/hdb3 showopts The device/file named in the resume= entry is a swap partition/file. If it doesn't exist, or isn't a properly formatted swap device, you will get a failure. You can eliminate the resume forever by removing that option from the kernel options, but you give up the ability to suspend-to-disk if you do that. -- The best way to accelerate a computer running Windows is at 9.81 m/s² -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Wednesday, January 17, 2007 @ 3:25 PM, I wrote:

Well, I've done some more digging on this problem and, though I have found out some things, I am still no closer to solving the problem than before.

1) The huge amount of lines in /var/log/messages seems to be unrelated to why I'm getting an fsck every time I boot. There are hundreds of lines in there that look like this one --

Linux kernel: SFW2-IN-ACC-RELATED IN eth0 OUT=MAC=00:08:74:24:85:82:00:04:5A:0f:18:07:08:00 SRC=128.61.111.11 DST=192.168.1.102 LEN=529 TOS=0x00 PREC=0x00 TTL=64 ID=238936 WINDOW=1716 RES=0x00 ACK FIN URGP=0 OPT (long hex number here)

The DST address is my machine's internal address. I checked several of the SRC addresses and they seemed to all be ZEN/YOU mirrors. So, this would appear to be some ZEN glitch (if so, what's new). I wouldn't think this would have anything to do with my getting the fscks every time I boot.

2) When I finally scrolled back across thousands of lines of the above and got to the beginning of the startup process, there were no messages that indicated any type of problem, disk related or otherwise.

3) I tried "tune2fs -c 99 /dev/hda2" and it came back and said "Setting maximal mount count to 99". However, I still get an fsck every time I boot up.

So, at this point I'm stumped. There doesn't seem to be any error message coming out and yet it just automatically does an fsck every time I boot. I'm going to look at some of the other files in /var/log to see if I can find one with some sort of message in it that would point me to why this is happening. Don't know what else to do at this point.

Greg Wallace

I just noticed the following lines ahead of the fsck in the log -- Invoking userspace resume from /dev/hda1 resume: Could not stat configuration file resume: libgcrypt version: 1.2.3 resume: Could not read the image Invoking in-kernel resume from /dev/hda1 <-- swap Waiting for device /dev/hda2 to appear ok fsck... Could the fact that it's trying to do a resume and is unable to be the cause of the problem? I really don't need resume from disk anyway. How can I turn that off? I'll start looking for that setting under /etc/sysconfig Editor. Greg W -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Greg Wallace wrote:

On Wednesday, January 17, 2007 @ 3:25 PM, I wrote:

...

Well, I've done some more digging on this problem and, though I have found out some things, I am still no closer to solving the problem than before.

...

1) The huge amount of lines in /var/log/messages seems to be unrelated to why I'm getting an fsck every time I boot. There are hundreds of lines in there that look like this one --

...

Linux kernel: SFW2-IN-ACC-RELATED IN eth0 OUT=MAC=00:08:74:24:85:82:00:04:5A:0f:18:07:08:00 SRC=128.61.111.11 DST=192.168.1.102 LEN=529 TOS=0x00 PREC=0x00 TTL=64 ID=238936 WINDOW=1716 RES=0x00 ACK FIN URGP=0 OPT (long hex number here)

...

The DST address is my machine's internal address. I checked several of the SRC addresses and they seemed to all be ZEN/YOU mirrors. So, this would appear to be some ZEN glitch (if so, what's new). I wouldn't think this would have anything to do with my getting the fscks every time I boot.

...

2) When I finally scrolled back across thousands of lines of the above and got to the beginning of the startup process, there were no messages that indicated any type of problem, disk related or otherwise.

...

3) I tried "tune2fs -c 99 /dev/hda2" and it came back and said "Setting maximal mount count to 99". However, I still get an fsck every time I boot up.

...

So, at this point I'm stumped. There doesn't seem to be any error message coming out and yet it just automatically does an fsck every time I boot. I'm going to look at some of the other files in /var/log to see if I can find one with some sort of message in it that would point me to why this is happening. Don't know what else to do at this point.

...

Greg Wallace

I just noticed the following lines ahead of the fsck in the log --

Invoking userspace resume from /dev/hda1 resume: Could not stat configuration file resume: libgcrypt version: 1.2.3 resume: Could not read the image Invoking in-kernel resume from /dev/hda1 <-- swap Waiting for device /dev/hda2 to appear ok fsck...

Could the fact that it's trying to do a resume and is unable to be the cause of the problem? I really don't need resume from disk anyway. How can I turn that off? I'll start looking for that setting under /etc/sysconfig Editor.

Greg W

I had a similiar problem a long time ago. You might want to bring it do run level 3, single user and do a manual fsck with full checking and repair. Sometimes, the fsck that is started up in the standard system is not doing an adequate repair. This may repair the disk correctly. After that you can reboot the system with a sync before. This may work but then may not. By the way, are you running ext2 as your file system? -- Joseph Loo jloo@acm.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On 1/17/07, Greg Wallace wrote:

Sounds like a good idea. It's been a long time since I've worked in other than graphical mode. I tried CTRL-ALT-PF2 and then tried to unmount the partition but it says it's busy. I've rarely worked in non-graphical mode. Can you suspend the graphical mode via some set of keys, unmount the file system and do an fsck, or do you have to boot up in non-graphical mode to begin with. If the latter, how is that done. I seem to recall there being a way to hit some key and enter a number to tell the system what level to boot to, but I can't remember the details.

CTRL-ALT-F2 # su # init 3 This will switch to runlevel 3 w/o need to reboot. But, if the partition in case is the one, which is used for / , then this will not help. Boot with the install cd/dvd, and select maintenace mode or whatever it is called. This will boot from the install media, and will not mount the root filesystem. You will be able to fsck from there. -- Svetoslav Milenov (Sunny) Even the most advanced equipment in the hands of the ignorant is just a pile of scrap. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Wed, Jan 17, 2007 at 09:14:14PM -0600, Greg Wallace wrote:

I said fix and it came back saying the volume was clean. I then rebooted normally and it still ran an fsck. So, I went back and booted from the rescue disk again, ran another e2fsck, and got that same message about the superblock last write time being in the future. I again fixed it, rebooted, and it still runs an fsck. Really strange.

<sigh> Fix your clock. Let me guess... you left the hardware clock on local time? -- Marc Wilson | Computers can figure out all kinds of problems, msw@cox.net | except the things in the world that just don't | add up. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Wed, Jan 17, 2007 at 11:05:08PM -0600, Greg Wallace wrote:

Not sure what you mean by "hardware clock".

Uh, the one built into the computer? The one that when you installed, you were asked whether it was set to local time, or UTC? -- Marc Wilson | Veni, Vidi, Visa. msw@cox.net | -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Wednesday 2007-01-17 at 23:05 -0600, Greg Wallace wrote:

...

<sigh> Fix your clock. Let me guess... you left the hardware clock on local time?

Not sure what you mean by "hardware clock".

The one you see in the bios setup. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFFr1ATtTMYHG2NR9URAnyhAJ9sY6evR+OihyRCuVjqYbr2gjIEdQCffu8X VemUWnE6HOFgGaGOtTqfSaw= =AK4o -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

* Greg Wallace [01-18-07 14:38]:

On Thursday, January 18, 2007 @ 4:47 AM, Carlos Robinson wrote:

...

The Wednesday 2007-01-17 at 23:05 -0600, Greg Wallace wrote:

...

...

...

<sigh> Fix your clock. Let me guess... you left the hardware clock on local time?

Not sure what you mean by "hardware clock".

...

The one you see in the bios setup.

I'm running a Dell Optiplex GX260. In the BIOS, the only setting for the clock is a digital clock where you adjust the time manually. There is no option as far as local time vs any other type of time.

You mean that you cannot set the clock in the cmos? You are not thinking here. Local time is the time you see over your shoulder on the *wall* clock. UTF is the time at the zero meridian, used to be universal mean time and/or zulu time. YOU are the one who determines what time is in the cmos. -- Patrick Shanahan Registered Linux User #207535 http://wahoo.no-ip.org @ http://counter.li.org HOG # US1244711 Photo Album: http://wahoo.no-ip.org/gallery2 OpenSUSE Linux http://en.opensuse.org/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Thursday 18 January 2007 14:10, Greg Wallace wrote:

Duh, where is the cmos? As far as local time vs UTF, I understand the difference quite well. It's just that I don't see any option of using it anywhere I have looked (bios and right clicking the clock in SUSE). Is this cmos a setting that is somewhere else to be found?

Thanks, Greg Wallace

This is all FYI. BIOS=CMOS=hardware clock. Right-click the clock in KDE's systray and choose 'Adjust Date & Time...", enter the root password, click the Configure button and I end up in NTP configuration in openSUSE 10.2...?!? Not what I expected. NTP is the way to go if you have broadband. Keeps your system on time, all the time. I advocate you use us.pool.ntp.org in the USA (substitute your 2 letter country code elsewhere) because it keeps it geographically local. Choosing the 'Use Random Servers from pool.ntp.org' randomly chooses from around the globe. Netiquette of NTP says use a Stratum 2 or 3 server geographically close and as little as possible. man ntpd has good intro info for NTP. Right-click on KDE's systray clock used to put you into YaST, System, Date & Time settings. That's where the 'Hardware Clock Set To' offers 2 choices of either 'Local Time' or 'UTC'. This is the same screen you see during an Install of openSUSE. I always choose 'Local Time' because I occasionally boot into Windows. If my system(s) were set to 'UTC', Windows would reliably mess up the BIOS=CMOS=hardware clock and then SUSE's time would be off when I rebooted to it. I have no idea if this has been resolved. This is also mentioned in the side-bar help area of YaST, System, Date & Time. My system automatically adjusts for DST while set to 'Local Time'. So where it says "If the hardware clock is set to UTC, your system can switch from standard time to daylight saving time and back automatically." is wrong because mine always gets it right even on localtime. YMMV, Stan -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Thursday 2007-01-18 at 17:58 -0600, Greg Wallace wrote:

...

This is also mentioned in the side-bar help area of YaST, System, Date & Time.

Ok, now in there there IS an option to select UTC, so I changed it. However, my time is now shown in UTC. How can I get it to show local time on my clock when UTC is selected. If that isn't possible, then I'll have to change it back.

If you set it to UTC it will obviously show UTC time there. But the time displayed in your kde bar (for instance) doesn't change, it will keep correcct local time. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFFsAw8tTMYHG2NR9URAgaZAJ9P8ZFt/rsTa88yNUJqTWgasjSgDQCdGnL0 ZPFRD0/YkE32INDYQKWh4SA= =SE5N -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Thursday 2007-01-18 at 14:10 -0600, Greg Wallace wrote:

...

...

...

...

Not sure what you mean by "hardware clock".

...

The one you see in the bios setup.

I'm running a Dell Optiplex GX260. In the BIOS, the only setting for the clock is a digital clock where you adjust the time manually. There is no option as far as local time vs any other type of time.

You surely can adjust that clock! You decide if you type your local time, or you type GMT/UTC time. And only you know what you typed.

...

You mean that you cannot set the clock in the cmos? You are not thinking here. Local time is the time you see over your shoulder on the *wall* clock. UTF is the time at the zero meridian, used to be universal mean time and/or zulu time. YOU are the one who determines what time is in the cmos.

Duh, where is the cmos?

It's another name for the bios. It's the same thing.

As far as local time vs UTF, I understand the difference quite well. It's just that I don't see any option of using it anywhere I have looked (bios and right clicking the clock in SUSE). Is this cmos a setting that is somewhere else to be found?

You set up your clock as you like, then you tell Yast what you did (clock setting somewhere). - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFFsAuBtTMYHG2NR9URAokUAJ9Nb+f0dmXeHi0xGA1AHZgfQsW3TgCfQPZ9 jjIdWd1miqxiyt8okTE1+KY= =lE35 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Thursday, January 18, 2007 @ 1:54 PM, Darryl Gregorash wrote:

...

<snip>

I'm running a Dell Optiplex GX260. In the BIOS, the only setting for the clock is a digital clock where you adjust the time manually. There is no option as far as local time vs any other type of time. Correct. The BIOS clock doesn't know about anything beyond itself. It

On 2007-01-18 13:26, Greg Wallace wrote: probably doesn't know too much about that either.

If you want to set your clock to UTC, add 6 hours to your local time (since you are in Central Time). Otherwise, set it to local time.

Thanks. This entire clock conversation got started when someone indicated that my being on local time was the cause of my fsck running every time I boot up. Somehow, I don't think advancing my clock 6 hours (I think that's how far behind GMT I am here in the Central zone) will fix that problem. It doesn't really seem logical that that is what is causing it, but maybe I'm wrong. Greg Wallace -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Thursday 2007-01-18 at 14:15 -0600, Greg Wallace wrote:

Thanks. This entire clock conversation got started when someone indicated that my being on local time was the cause of my fsck running every time I boot up. Somehow, I don't think advancing my clock 6 hours (I think that's how far behind GMT I am here in the Central zone) will fix that problem. It doesn't really seem logical that that is what is causing it, but maybe I'm wrong.

I don't think so either. If your time is displaying correctly in your system, then the clock is not the problem, IMO. There is some funny error in the filesystem, but I don't know what it really means. Do you perchance have a /forcefsck file? - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFFsA1TtTMYHG2NR9URAoDAAJwJEMuOJhTVABulVuff2MCF+N5RswCfZ3vf BhMe6E6Utru/wF1ruSv/Tyg= =D0sg -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Thu, Jan 18, 2007 at 02:15:51PM -0600, Greg Wallace wrote:

Thanks. This entire clock conversation got started when someone indicated that my being on local time was the cause of my fsck running every time I boot up. Somehow, I don't think advancing my clock 6 hours (I think that's how far behind GMT I am here in the Central zone) will fix that problem. It doesn't really seem logical that that is what is causing it, but maybe I'm wrong.

You're wrong. Where else did you think the "furure" errors were coming from? -- Marc Wilson | "I'm willing to sacrifice anything for this cause, msw@cox.net | even other people's lives" -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Greg Wallace wrote:

On Thursday, January 18, 2007 @ 4:47 AM, Carlos Robinson wrote:

...

The Wednesday 2007-01-17 at 23:05 -0600, Greg Wallace wrote:

...

...

...

<sigh> Fix your clock. Let me guess... you left the hardware clock on local time?

Not sure what you mean by "hardware clock".

...

The one you see in the bios setup.

...

- -- Cheers, Carlos E. R.

I'm running a Dell Optiplex GX260. In the BIOS, the only setting for the clock is a digital clock where you adjust the time manually. There is no option as far as local time vs any other type of time.

Greg Wallace

When you load the system up, you have the option of setting either local time or UTC time. I believe that is what he is saying. If the superblock is always saying it is in the future, make sure you do a sync before you do a reboot. It is very rare, but it is just to be careful. Try to see if it is having the same problem. If you are still having the problem, there is a good chance that the sector is going bade. You might want to run soething like Steve Gibson Spinrite. In any case, I strongly suggest you make a backup of your system, as there is a good chance that your drive is beginning to die. -- Joseph Loo jloo@acm.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Greg Wallace wrote:

On Thursday, January 18, 2007 @ 10:06 PM, Joseph Loo wrote:

...

Greg Wallace wrote:

...

On Thursday, January 18, 2007 @ 4:47 AM, Carlos Robinson wrote:

...

The Wednesday 2007-01-17 at 23:05 -0600, Greg Wallace wrote:

...

...

...

<sigh> Fix your clock. Let me guess... you left the hardware clock on local time?

Not sure what you mean by "hardware clock".

...

The one you see in the bios setup.

...

- -- Cheers, Carlos E. R.

I'm running a Dell Optiplex GX260. In the BIOS, the only setting for the clock is a digital clock where you adjust the time manually. There is no option as far as local time vs any other type of time.

Greg Wallace

When you load the system up, you have the option of setting either local time or UTC time. I believe that is what he is saying.

...

If the superblock is always saying it is in the future, make sure you do a sync before you do a reboot. It is very rare, but it is just to be careful. Try to see if it is having the same problem.

...

If you are still having the problem, there is a good chance that the sector is going bade. You might want to run soething like Steve Gibson Spinrite. In any case, I strongly suggest you make a backup of your system, as there is a good chance that your drive is beginning to die.

...

--

...

Joseph Loo jloo@acm.org

I changed to UTC and the message about the time being in the future went away. However, I still get a message saying it's doing an fsck. The difference, though, is that now, rather than taking several minutes to go by that message it goes by it instantly, so I don't know if it's really doing an fsck anymore or not.

Greg Wallace

It has been a long time since I worked with fsck. I could be wrong, but I believe it always does a fsck to check for file system consistency. -- Joseph Loo jloo@acm.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On 2007-01-19 18:47, Greg Wallace wrote:

On Friday, January 19, 2007 @ 6:29 PM, Joseph Loo wrote:

<snip>

...

It has been a long time since I worked with fsck. I could be wrong, but I believe it always does a fsck to check for file system consistency.

That certainly seems to be the case now, and maybe for some time now and I just hadn't noticed. But I don't think that's been the case for too many

I may have mentioned in another note that this is likely due to the introduction of journalled file systems. I believe the situation is roughly this: Re-running the journal on boot allows the system to keep the file system in better order. Before that happens, however, a consistency check must be done on the file system. -- The best way to accelerate a computer running Windows is at 9.81 m/s² -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On 2007-01-20 06:39, Thomas Hertweck wrote:

Darryl Gregorash wrote:

...

... Re-running the journal on boot allows the system to keep the file system in better order. Before that happens, however, a consistency check must be done on the file system.

Journal replaying itself is done to ensure the consistency of the filesystem. Since the journal contains a chronological log of recent metadata changes, it's able to simply check the portions that have recently been modified - which is a matter of seconds. It does not re-order anything. After this step, the filesystem is usually marked as clean (unless something goes wrong and an fsck might be forced).

The reason I wrote that way is because my boot.msg contains the following: Checking file systems... fsck 1.36 (05-Feb-2005) Reiserfs super block in block 16 on 0x349 of format 3.6 with standard journal Blocks (total/free): 2008112/1231671 by 4096 bytes Filesystem is clean Replaying journal.. <etc> Am I correct in viewing the whole thing, super block plus tree structure, as the "file system?" It would seem that first there is a check for a valid super block, then the tree is checked for consistency, and only then is the journal replayed. (Other journalling f ilesystems may vary in specifics, of course.) -- The best way to accelerate a computer running Windows is at 9.81 m/s² -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On 2007-01-22 13:09, Thomas Hertweck wrote:

<snip>

I don't know details about ReiserFS - this FS has been banned from all our systems a long time ago. I know a bit about ext3 and xfs though.

Interesting. I'd be interested in knowing why, if you can share that information. I'm using a Reiser-only installation right now, but it requires a rather large journal (at least, that is what the installation told me when I tried to create a small /boot partition, about 15 MB or so -- minimum partition size was stated at just over 100 MB). Ext3 seems like a good alternative; I've never used xfs at all.

When the filesystem is marked as clean, then there is usually no need to do an fsck or to replay the journal - the filesystem should be in a

I thought you had to run fsck to determine if the filesystem is clean? As for replaying the journal, that seems to be a constant with Reiser. What you say definitely makes sense, but seems to be contradicted in practice, at least insofar as my log files tell me. -- The best way to accelerate a computer running Windows is at 9.81 m/s² -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Thomas Hertweck wrote:

Darryl Gregorash wrote:

...

On 2007-01-22 13:09, Thomas Hertweck wrote:

...

<snip>

I don't know details about ReiserFS - this FS has been banned from all our systems a long time ago. I know a bit about ext3 and xfs though.

Interesting. I'd be interested in knowing why, if you can share that information.

We have used ReiserFS when it became known and when it was available in Linux distributions. However, at that time we faced lots of weird problems that were caused by the filesystem (yes, these were problems with the FS and not the hardware) and reiserfsck was not working very well, i.e. if something went wrong, we had situation where reiserfsck made things much worse instead of repairing the FS. We then decided that this is certainly not a filesystem we would like to use in a production environment. Since we usually have to deal with large systems (tens or hundreds of TiB) and (very) large files, we decided to try xfs (which we already knew from our SGI servers) - xfs was designed and optimized for large filesystems and large files right from the beginning. It turned out to be a very good decision and, thus, there was no need to come back and try ReiserFS. On thin clients or laptops and/or dual-boot machines, we are using ext3. It's a bit safer when it comes to losing data at power cuts (which is due to the way the journalling works) - our desktops have no UPS. Furthermore, there exist tools to access ext partitions from Windows OS.

Sorry, but that just sounds too bizarre to me - we've been using reiserfs in our data center for years, since it's the default filesystem on suse linux. It's been rock solid here, and is also the fastest journaling filesystem we've found. I understand that reiserfs maintenance going forward may slow down, so we'll have to eventually settle on a different filesystem, but there's no reason to suddenly change all of our stable systems on a political whim. Hopefully some other filesystem will be able to fill the gap - possibly ext4, or xfs. Joe -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Tuesday 2007-01-23 at 00:46 -0600, Darryl Gregorash wrote:

...

I don't know details about ReiserFS - this FS has been banned from all our systems a long time ago. I know a bit about ext3 and xfs though.

Interesting. I'd be interested in knowing why, if you can share that information.

You can find many people against, and many for it.

I'm using a Reiser-only installation right now, but it requires a rather large journal (at least, that is what the installation told me when I tried to create a small /boot partition, about 15 MB or so -- minimum partition size was stated at just over 100 MB).

Not exactly a large journal. But yes, there is a minimun size for a reiserfs parttition. I use ext2 for /boot.

Ext3 seems like a good alternative; I've never used xfs at all.

xfs is very good for large files. Some people recomend it for the /home directory. I use it for DVDs as well.

...

When the filesystem is marked as clean, then there is usually no need to do an fsck or to replay the journal - the filesystem should be in a

I thought you had to run fsck to determine if the filesystem is clean? As for replaying the journal, that seems to be a constant with Reiser. What you say definitely makes sense, but seems to be contradicted in practice, at least insofar as my log files tell me.

It is normal to do a fast test of reiserfs partitions during boot, and AFAIK, it run always. I'm not sure for ext3, but that could be also the case. XFS does also run a quick check when mounting. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFFtqdftTMYHG2NR9URAjEmAJ9CAUksdBVNvTMUCYOF5ZCfZ0HGCwCePFYE WX7YZj1CnRpwrnFC9ym0XSs= =/PTE -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On 2007-01-17 19:08, Greg Wallace wrote:

<snip> Can you suspend the graphical mode via some set of keys, unmount the file system and do an fsck, or do you have to boot up in non-graphical mode to

Neither is a viable option, because you will be unable to umount essential parts of the filesystem, like /. Running fsck on a mounted file system is, as they would say in Oceania[1], "double plus ungood". Boot to the rescue system on your installation CD/DVD. The filesystem there is just a RAMdisk, and nothing else is mounted. You can fsck every partition in your system without worry. -- [1] This is 2007, don't tell me you haven't read Orwell?!! -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Thursday, January 18, 2007 @ 4:44 AM, Carlos Robinson wrote:

The Wednesday 2007-01-17 at 16:42 -0600, Greg Wallace wrote:

...

I just noticed the following lines ahead of the fsck in the log --

Invoking userspace resume from /dev/hda1 resume: Could not stat configuration file resume: libgcrypt version: 1.2.3 resume: Could not read the image Invoking in-kernel resume from /dev/hda1 <-- swap Waiting for device /dev/hda2 to appear ok fsck...

Could the fact that it's trying to do a resume and is unable to be the cause of the problem?

No.

Better post the following lines to that "fsck..." :-)

- -- Cheers, Carlos E. R.

Carlos: Here's all of the lines from the start-up log beginning right before the fsck and ending with the mount of the file system. Waiting for device /dev/hda2 to appear: OK fsck 1.39 (29-May-2006) [/bin/fsck/ext3 (1) -- /] fsck.ext3 -a -CO /dev/hda2 / (/dev/hda2): Superblock last write time is in the future. FIXED / (/dev/hda2) clean, 34... files, 89... blocks fsck succeeded. Mounting root device read write Mounting root /dev/hda2 I don't understand the part about last write time being in the future. Every time I boot I get that same message. When I booted from the DVD and ran e2fsck I also got it. I followed with another e2fsck right afterward just to see if it went away and it did. But, when I booted again normally it showed right back up again. Do you think that's what's causing the fsck? I. e., when the system goes to mount the file it sees this future date and that triggers an fsck. If so, I have no earthly idea what is causing that date problem. Greg Wallace -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Greg Wallace wrote:

...

Here's all of the lines from the start-up log beginning right before

On Thursday, January 18, 2007 @ 5:51 AM, Joe Morris wrote: the

...

fsck and ending with the mount of the file system.

Waiting for device /dev/hda2 to appear: OK fsck 1.39 (29-May-2006) [/bin/fsck/ext3 (1) -- /] fsck.ext3 -a -CO /dev/hda2 / (/dev/hda2): Superblock last write time is in the future. FIXED / (/dev/hda2) clean, 34... files, 89... blocks fsck succeeded. Mounting root device read write Mounting root /dev/hda2

I don't understand the part about last write time being in the future. Every time I boot I get that same message. When I booted from the DVD and ran e2fsck I also got it. I followed with another e2fsck right afterward just to see if it went away and it did. But, when I booted again normally it showed right back up again. Do you think that's what's causing the fsck?

Yes, definitely. So to recap, you do not have a fsck problem, you have a time problem causing it to update the superblock (my guess is the writing of the dirty bit) with the wrong date, which causes it to fail the initial fsck and causing it to run to fix that problem. When you shutdown, it updates the hardware clock from the system clock (probably not the problem) and should mark the filesystem as cleanly shutdown. Something is writing the wrong date. Since you had a syslog-ng.conf.rpmnew file, is it possible there are other config files you have not updated since this was obviously an update install?

-- Joe Morris Registered Linux user 231871 running openSUSE 10.2 x86_64

Switching to UTC caused that date message to go away! Now, the only thing that comes out is - Waiting for device /dev/hda2 to appear: OK fsck 1.39 (29-May-2006) [/bin/fsck/ext3 (1) -- /] fsck.ext3 -a -CO /dev/hda2 / (/dev/hda2) clean, 34... files, 89... blocks fsck succeeded. Mounting root device read write Mounting root /dev/hda2 So if that date problem is cleared up, why do I still get an fsck every time I boot? Here is a list of all of the rpmnews on my system. /etc/cups/printers.conf.rpmnew /etc/init.d/smbfs.rpmnew /etc/inittab.rpmnew /etc/krb5.conf.rpmnew /etc/ldap.conf.rpmnew /etc/localtime.rpmnew /etc/magic.rpmnew /etc/networks.rpmnew /etc/nsswitch.conf.rpmnew /etc/opt/kde3/share/config/k3b/k3bsetup.rpmnew /etc/opt/kde3/share/config/khelpcenterrc.rpmnew /etc/opt/kde3/share/config/kioslaverc.rpmnew /etc/postfix/main.cf.rpmnew /etc/samba/lmhosts.rpmnew /etc/samba/smb.conf.rpmnew /etc/samba/smb.conf.rpmnew~ /etc/samba/smbfstab.rpmnew /etc/samba/smbpasswd.rpmnew /etc/sane.d/dll.conf.rpmnew /etc/security/pam_unix2.conf.rpmnew /etc/sudoers.rpmnew /etc/syslog-ng/syslog-ng.conf.rpmnew /etc/X11/qtrc.rpmnew /etc/xinetd.d/swat.rpmnew /opt/kde3/share/config/kdm/kdmrc.rpmnew /etc/udev/rules.d/30-net_persistent_names.rules.rpmnew /usr/sbin/useradd.local.rpmnew Greg Wallace -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Thursday, January 18, 2007 @ 10:51 PM, Darryl Gregorash wrote:

On 2007-01-18 20:42, Greg Wallace wrote:

...

<snip>

So if that date problem is cleared up, why do I still get an fsck every time I boot? Here is a list of all of the rpmnews on my system.

I suspect you're getting a fsck each time you boot because you're using a journalled filesystem. The same happens here, with Reiserfs. It is a simple consistency check of the system.

Ok. Sounds good. It races right by it during the boot process, so the simple fact that those messages come out doesn't bother me. I consider the problem solved. Thanks again, Greg Wallace -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Friday, January 19, 2007 @ 5:45 AM, Joe Morris wrote:

Greg Wallace wrote:

...

Switching to UTC caused that date message to go away!

So you must have the setting that said your bios clock was set to utc, yet before it was actually set to localtime. They only need to agree.

I'm still confused about this. My BIOS doesn't have the ability to use UTC vs Local. Plus, I haven't changed my clock configuration since 8.1 and have never had this problem before, so I guess it's something new with 10.2. Anyway, it's working now so I guess that's all that matters.

...

Now, the only thing that comes out is -

Waiting for device /dev/hda2 to appear: OK fsck 1.39 (29-May-2006) [/bin/fsck/ext3 (1) -- /] fsck.ext3 -a -CO /dev/hda2 / (/dev/hda2) clean, 34... files, 89... blocks fsck succeeded. Mounting root device read write Mounting root /dev/hda2

So if that date problem is cleared up, why do I still get an fsck every time I boot?

It is an automatic check to make sure no filesystem needs more extensive checking.

Ok. I always thought that the only time an fsck was invoked was when the number of mounts you had done equaled the setting for that. In the past when an fsck occurred it was always accompanied by a message saying something like "number of mounts since last check exceeded. fsck forced", or something to that effect. Otherwise, I don't ever seeing the words fsck in the log at all. Now it's in the log every time I boot. Maybe it was there before and I just didn't notice it.

...

Here is a list of all of the rpmnews on my system.

/etc/cups/printers.conf.rpmnew /etc/init.d/smbfs.rpmnew /etc/inittab.rpmnew /etc/krb5.conf.rpmnew /etc/ldap.conf.rpmnew /etc/localtime.rpmnew /etc/magic.rpmnew /etc/networks.rpmnew /etc/nsswitch.conf.rpmnew /etc/opt/kde3/share/config/k3b/k3bsetup.rpmnew /etc/opt/kde3/share/config/khelpcenterrc.rpmnew /etc/opt/kde3/share/config/kioslaverc.rpmnew /etc/postfix/main.cf.rpmnew /etc/samba/lmhosts.rpmnew /etc/samba/smb.conf.rpmnew /etc/samba/smb.conf.rpmnew~ /etc/samba/smbfstab.rpmnew /etc/samba/smbpasswd.rpmnew /etc/sane.d/dll.conf.rpmnew /etc/security/pam_unix2.conf.rpmnew /etc/sudoers.rpmnew /etc/syslog-ng/syslog-ng.conf.rpmnew /etc/X11/qtrc.rpmnew /etc/xinetd.d/swat.rpmnew /opt/kde3/share/config/kdm/kdmrc.rpmnew /etc/udev/rules.d/30-net_persistent_names.rules.rpmnew /usr/sbin/useradd.local.rpmnew

You should really go through every one of those files and either replace the orig with the newer one, or at least deal with all those config files. I see /etc/localtime is one of those files. That was probably your time problem's root. If you are going to update your install, you need to finish it by checking and resolving these config files.

Yeah. I just recently (a few weeks ago) became aware of what it meant to have RPMNEW files. Should I go through them one by one and see if the date of the RPMNEW is later than the date of the RPM (else, the RPMNEW was created in an earlier install and the RPM would actually be the latest version, right?). For those where the RPMNEW is the latest, is it pretty safe to just swap them in for the regular rpm file? I guess I'd want to do them a few at a time so that if I foul up my system I'll know which one to swap back.

-- Joe Morris Registered Linux user 231871 running openSUSE 10.2 x86_64

Greg Wallace -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Friday, January 19, 2007 @ 6:50 PM, Carlos Robinson wrote:

The Friday 2007-01-19 at 15:03 -0600, Greg Wallace wrote:

...

...

So you must have the setting that said your bios clock was set to utc, yet before it was actually set to localtime. They only need to agree.

I'm still confused about this. My BIOS doesn't have the ability to use UTC vs Local.

Ffffssssss.....

No bios has that ability. No one has it. We did not say that. You are confusing what we write.

The bios clock (aka cmos clock aka hardware clock) does not know and does not care if the time it keeps is local or utc or martian. Only the operation system knows about such subtleties. Ok?

...

Plus, I haven't changed my clock configuration since 8.1 and have never had this problem before, so I guess it's something new with 10.2.

No, it's not new. It has always been the same way, even before 8.x

Switching to UTC caused the long running fscks to go away, so it seems to me that there was a direct correlation between the two. I had never changed from local to utc or vice versa in the past, so something must have changed with 10.2 to cause this connection. This is, of course, pure speculation, but it seems to fit with what was occurring.

...

Anyway, it's working now so I guess that's all that matters.

Right.

...

...

It is an automatic check to make sure no filesystem needs more extensive checking.

Ok. I always thought that the only time an fsck was invoked was when the number of mounts you had done equaled the setting for that. In the past when an fsck occurred it was always accompanied by a message saying something like "number of mounts since last check exceeded. fsck forced",

That's the long check, and the one for ext2 types.

I didn't realize there was more than one kind of fsck. Good information. Thanks.

...

or something to that effect. Otherwise, I don't ever seeing the words fsck in the log at all. Now it's in the log every time I boot. Maybe it was there before and I just didn't notice it.

And maybe the log handling has been improved and now it is being logged ;-)

Perhaps that's it.

...

...

You should really go through every one of those files and either replace the orig with the newer one, or at least deal with all those config files. I see /etc/localtime is one of those files. That was probably your time problem's root. If you are going to update your install, you need to finish it by checking and resolving these config files.

Yeah. I just recently (a few weeks ago) became aware of what it meant to have RPMNEW files. Should I go through them one by one and see if the date of the RPMNEW is later than the date of the RPM (else, the RPMNEW was created in an earlier install and the RPM would actually be the latest version, right?). For those where the RPMNEW is the latest, is it pretty safe to just swap them in for the regular rpm file?

Sometimes yes, sometimes no. :-p

...

I guess I'd want to do them a few at a time so that if I foul up my system I'll know which one to swap back.

You have to look at the dates, yes, and at what they contain, choose which one is the best one, and activate it. Or, activate the new one with your modifications from the original file. I usually move the original somewhere else so I know what I changed.

Yeah. I'm doing a full system backup and creating a log of which ones I'm changing in case something breaks. Then I'm going to go through one by one and examine them, renaming them to RPMOLD if they have been bypassed by later updates of the file and swapping the RPMNEWs in for the regular RPMs if the date on the RPMNEW is more current. If something fouls up, I can always recover from the backup.

- -- Cheers, Carlos E. R.

Thanks, Greg Wallace -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Friday, January 19, 2007 @ 8:56 PM, Joe Morris wrote:

Greg Wallace wrote:

...

Switching to UTC caused the long running fscks to go away, so it seems to me that there was a direct correlation between the two. There was. I had never changed from local to utc or vice versa in the past, so something must have changed with 10.2 to cause this connection. It was the fact you had never checked /etc/localtime.rpmnew This is, of course, pure speculation, but it seems to fit with what was occurring.

AFAICT, your BIOS clock was set to localtime

Here we go again with the BIOS clock being set to localtime. I've already been admonished on this topic and told that the BIOS clock is just set to a time, period and that localtime vs UTC only gets involved in the operating system. Then I hear see these types of notes and, of course, it's confusing.

, your setting was set to UTC,

SUSE was set to use local time when the problem was occurring.

and when it boot the time stamp for the hard drive was in the future as compared to the BIOS clock

I'm still confused here. Because I was using local time, I don't see how a UTC time was getting written to the hard drive. I guess, even though I was using local time, the process that wrote the time to the hard drive was using UTC all on its on; else how to explain it.

, causing it to be marked as inconsistent and an fsck run to fix the inconsistency. The problem could have been avoided by dealing with your localtime file or setting it to what your BIOS clock was set to in Yast. You set your clock to agree with your settings instead of vice versa, which also works.

PS Please learn to quote only the necessary parts of previous messages. They are much easier to read.

-- Joe Morris Registered Linux user 231871 running openSUSE 10.2 x86_64

-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Saturday, January 20, 2007 @ 9:51 AM, Carlos Robinson wrote:

The Friday 2007-01-19 at 21:57 -0600, Greg Wallace wrote:

...

...

AFAICT, your BIOS clock was set to localtime

Here we go again with the BIOS clock being set to localtime. I've already been admonished on this topic and told that the BIOS clock is just set to a time, period and that localtime vs UTC only gets involved in the operating system. Then I hear see these types of notes and, of course, it's confusing.

And here you go again. X-)

Yes, the bios clock can be set to local time, or to utc time, or to martian time, or whatever. Doesn't matter. The only thing is that the bios doesn't know what type of time it is! That's all. You set it to local time: fine. You set it to UTC: fine. Martian: fine. But don't try to ask the bios which time it has, it knows nothing. It's _you_ who know which time it keeps.

So, _you_ set the bios to martian time, then _you_ tell the system to mind that the bios has martian time. Only make sure that it is correct.

Capishi? :-)

Ok. I see what you're saying, but there seemed to be some implication that I could resolve my getting a time error message and a full fsck by changing the BIOS clock to UTC. I still don't understand that. The problem was resolved by my simply changing the SUSE time to be UTC based instead of local based (I never adjusted the hardware clock). And I still don't understand how that was affecting fsck. I mean, if it's using the same clock from one boot up to the next, why would it think the time found in the file system was in the future every time? But that change did fix the problem, so, whatever works as they say.

...

...

PS Please learn to quote only the necessary parts of previous messages. They are much easier to read.

¡¡¡YES, PLEASE!!!

¡Remove unneeded text from emails!

- -- Cheers, Carlos E. R.

-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Saturday, January 20, 2007 @ 7:20 PM, Carlos E. R. wrote:

...

I still don't understand that. The problem was resolved by my simply changing the SUSE time to be UTC based instead of local based (I never adjusted the hardware clock).

No, but your linux will do that for you behind your back :-p

Ok. I checked my BIOS clock and it had indeed been advanced 6 hours to show UTC. Still don't see how that fixed the fsck problem, but it is working and, as you say, that's all that matters.

I know, I can tell you the exact script that does it. Believe me, it is changed.

Cheers, Carlos E. R.

Greg W -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Sunday, January 21, 2007 @ 2:56 PM, Carlos E. R. wrote: <snip>

Possibly the fsck routine was using the cmos clock to know the current time. Yeah, that's it. Look:

nimrodel:~ # ls /etc/init.d/boot.d/*localfs /etc/init.d/boot.d/*clock /etc/init.d/boot.d/K18boot.clock /etc/init.d/boot.d/S07boot.localfs /etc/init.d/boot.d/K19boot.localfs /etc/init.d/boot.d/S08boot.clock

The S07boot.localfs executes before /etc/init.d/boot.d/S08boot.clock, meaning that the fsck program runs before the system clock is properly set up.

Ok. I understand what you're saying, but that still means that the system had chosen to write a UTC date to the file system even though I was using local time on my machine (and always had been since 8.1). Seems like the date written to the file system should be the same as the one you are using in your system. I assume they always were in the past and that's why I never had this problem before. As a matter of fact, the fact that I can't use local time anymore without having an fsck run every time I boot is a bug. I would assume it's just a bug that somehow got introduced into my particular installation and is not widespread (haven't seen anyone else complain about it and surely someone else is using local time). At least converting to UTC fixed it without any noticeable difference on my system (except if I happen to be going into the bios and look at the date there). <snip>

- -- Cheers, Carlos E. R.

Greg W -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Saturday 2007-01-20 at 22:45 -0600, Darryl Gregorash wrote:

...

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Ok? :-)

Ie, trim the quoted text. I have left at the bottom some text for you to practice removing ;-) My, such subtlety.

Mind if I steal it? ;-)

Be my guest X'-) It was past betdtime, perhaps I wasn't in my most patient mode O:-) - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFFs9HZtTMYHG2NR9URAgfBAJ9YnljEma6MeYn13dYIwVaG2qfD7QCeN69Y SAnsFNakc12I4V0BqSs/XRU= =B63B -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Saturday, January 20, 2007 @ 9:44 AM, Carlos Robinson wrote:

The Friday 2007-01-19 at 20:13 -0600, Greg Wallace wrote:

...

Yeah. I'm doing a full system backup and creating a log of which ones I'm changing in case something breaks. Then I'm going to go through one by one and examine them, renaming them to RPMOLD if they have been bypassed by later updates of the file and swapping the RPMNEWs in for the regular RPMs if the date on the RPMNEW is more current. If something fouls up, I can always recover from the backup.

Don't rename, just move to some other dir. The rpmold name is also used sometimes, besides rpmnew. Read the man.

- -- Cheers, Carlos E. R.

Ok. I'll bear that in mind. I went through man rpm and couldn't find anything about this. I had done a Google earlier and found the following. It only mentions RPMNEW and RPMSAVE, not RPMOLD. http://www.redhat.com/archives/rhl-list/2003-December/msg04713.html Greg W -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org