On Fri, 16 Nov 2007 11:06:43 +1100, Doug Laidlaw wrote:
> Running Mdv 2008.

Me to.
$ cat /etc/release
Mandriva Linux release 2008.0 (Official) for i586


> I am getting Rpm database corruption.

Not me. and no one else has complained.
I have my system on 24/7.

> I have followed
> the procedure in the Errata, to rebuild the database, and it corrects the
> error, but I seem to need to do it on a daily basis. I have set up my
> system with cron to shut down after about 04:45 local,

Yeeouch, might be a better way.

> to allow the nightly housekeeping to proceed.

Now confused, how can you shut down system to let cron run.

> Perhaps I am truncating the database updating process?

I would give it first priority as a culprit.

> Should I perhaps go the next step, and install the development version of
> the packages they refer to?

Nothing like throwing, yet more variables/problems into the mix.

If I misunderstand what you want done, I would rename your cron job to
something like
xx_sys_shutdown and either place the code or links to it in
/etc/cron.daily
/etc/cron.weekly
/etc/cron.monthly
You need code in it to figure if is is first of month/week so as to allow
the weekly and monthly jobs to complete.

Bit Twister wrote:

>> to allow the nightly housekeeping to proceed.
>
> Now confused, how can you shut down system to let cron run.
>
Cart before horse. Root's cron shuts down system at 04:45 By then, all the
jobs in cron.daily, etc. seem to have finished. This scheme was quite
successful under 2007, then I went to running continuously. This may be
the first time I have tried it under 2008. (No written log of changes.)

Plainly the problem is arising at that point. Putting the shutdown in
cron.daily would fix it in the sequence, not by time. First, I will go
back to continuous. They say that it is the monitor that uses the power,
but I have a flat TFT display (can't even recall the usual name - my memory
is getting horrible,) and it blanks. More, I need to protect myself from
getting up during the night and playing computer.

Doug.

Doug Laidlaw writes:

>Running Mdv 2008. I am getting Rpm database corruption. I have followed
>the procedure in the Errata, to rebuild the database, and it corrects the
>error, but I seem to need to do it on a daily basis. I have set up my
>system with cron to shut down after about 04:45 local, to allow the nightly
>housekeeping to proceed. Perhaps I am truncating the database updating
>process?

I is certainly possible that your shutting down is causing trouble.
That time is when lots of housekeeping is done.Why not try not shutting
down and seeing what happens.


>Should I perhaps go the next step, and install the development version of
>the packages they refer to? I am one of those people who would be better
>off living with the problem than applying a buggy fix, particularly if those
>packages are not far off being released.

>Doug.

On Fri, 16 Nov 2007 14:47:23 +1100, Doug Laidlaw wrote:
>
> Plainly the problem is arising at that point. Putting the shutdown in
> cron.daily would fix it in the sequence, not by time.

You have no idea what time jobs may or may not complete.

My suggestion was to pick a name that would put it at the bottom of
the cron directory.

Why you ask?
$ cat /etc/crontab
SHELL=/bin/bash
PATH=/sbin:/bin:/usr/sbin:/usr/bin
MAILTO=root
HOME=/

# run-parts
01 * * * * root nice -n 19 run-parts --report /etc/cron.hourly
02 4 * * * root nice -n 19 run-parts --report /etc/cron.daily
22 4 * * 0 root nice -n 19 run-parts --report /etc/cron.weekly
42 4 1 * * root nice -n 19 run-parts --report /etc/cron.monthly


See, run-parts executes each script in the indicated directory.

$ ls /etc/cron.daily
aide.check logrotate mlocate.cron rpm tmpwatch
ckupdate_cron makewhatis.cron msec tetex.cron

If your script is named something like xx_shutdown,
it would run after tetex.cron

On Fri, 16 Nov 2007 14:47:23 +1100, Doug Laidlaw
wrote:

>More, I need to protect myself from
>getting up during the night and playing computer.

Why? :-)
--
PGP key ID 0xEB7180EC

Bit Twister wrote:

> On Fri, 16 Nov 2007 11:06:43 +1100, Doug Laidlaw wrote:
>> Running Mdv 2008.
>
> Me to.
> $ cat /etc/release
> Mandriva Linux release 2008.0 (Official) for i586
>
Does that mean that I don't have to observe Smart Questions from now on?

I try to follow SmartQuestions intelligently, and omit obviously irrelevant
setup details. In my post about the mounting question, I had been asked
previously for full details in another thread along similar lines. There
are still some users of older releases.

Doug.

Bit Twister wrote:

> On Fri, 16 Nov 2007 14:47:23 +1100, Doug Laidlaw wrote:
>>
>> Plainly the problem is arising at that point. Putting the shutdown in
>> cron.daily would fix it in the sequence, not by time.
>
> You have no idea what time jobs may or may not complete.
>
> My suggestion was to pick a name that would put it at the bottom of
> the cron directory.
>
> Why you ask?
> $ cat /etc/crontab
> SHELL=/bin/bash
> PATH=/sbin:/bin:/usr/sbin:/usr/bin
> MAILTO=root
> HOME=/
>
> # run-parts
> 01 * * * * root nice -n 19 run-parts --report /etc/cron.hourly
> 02 4 * * * root nice -n 19 run-parts --report /etc/cron.daily
> 22 4 * * 0 root nice -n 19 run-parts --report /etc/cron.weekly
> 42 4 1 * * root nice -n 19 run-parts --report /etc/cron.monthly
>
>
> See, run-parts executes each script in the indicated directory.
>
> $ ls /etc/cron.daily
> aide.check logrotate mlocate.cron rpm tmpwatch
> ckupdate_cron makewhatis.cron msec tetex.cron
>
> If your script is named something like xx_shutdown,
> it would run after tetex.cron

Tried that last night (too late at night.) Forgot to make the script
executable. Oh well, tonight is another night....

So, no shutdown; but the database is still corrupted again.

Doug.

On Sat, 17 Nov 2007 18:26:09 +1100, Doug Laidlaw wrote:
> Bit Twister wrote:
>
>> On Fri, 16 Nov 2007 11:06:43 +1100, Doug Laidlaw wrote:
>>> Running Mdv 2008.
>>
>> Me to.
>> $ cat /etc/release
>> Mandriva Linux release 2008.0 (Official) for i586
>>
> Does that mean that I don't have to observe Smart Questions from now on?

No, you should always observe Smart Questions.

I was showing you I had no urpm* woes on same release, for an i586.
Now if you had a 64 bit install, then you would have a possible suspect.

> I try to follow SmartQuestions intelligently, and omit obviously irrelevant
> setup details. In my post about the mounting question, I had been asked
> previously for full details in another thread along similar lines. There
> are still some users of older releases.

Yes, not to mention, 32/64 bit installs, of One Gnome/kde, Free, PowerPack

On Sat, 17 Nov 2007 18:28:45 +1100, Doug Laidlaw wrote:
>
> Tried that last night (too late at night.) Forgot to make the script
> executable. Oh well, tonight is another night....

You must be good at programming. I'd like to see that script.
Having it run in daily/weekly/monthly which may be running in parallel
is not a trivial task.


> So, no shutdown; but the database is still corrupted again.

Well, that would tend to rule out, your cron job, assuming you
remembered to disable the old one.

Time it install memtest86, boot it from boot menu, and let over night.
Guessing all that disk work, heats it up, cpu starts sweating, and you
have problems.

You need to rule out memory/cpu problem.

Bit Twister wrote:

> On Sat, 17 Nov 2007 18:28:45 +1100, Doug Laidlaw wrote:
>>
>> Tried that last night (too late at night.) Forgot to make the script
>> executable. Oh well, tonight is another night....
>
> You must be good at programming. I'd like to see that script.
> Having it run in daily/weekly/monthly which may be running in parallel
> is not a trivial task.
>
>
>> So, no shutdown; but the database is still corrupted again.
>
> Well, that would tend to rule out, your cron job, assuming you
> remembered to disable the old one.
>
> Time it install memtest86, boot it from boot menu, and let over night.
> Guessing all that disk work, heats it up, cpu starts sweating, and you
> have problems.
>
> You need to rule out memory/cpu problem.

I missed the earlier part of this thread, but if the OP did an upgrade,
rather than clean install, then there is a known bug:

http://qa.mandriva.com/show_bug.cgi?id=32547

It's a long complicated read, but from what I understand (with my very
limited technical knowledge) the latest identified potential "culprit"
seems to be a possible / filesystem corruption.

Jim

On Sat, 17 Nov 2007 16:33:26 +0000, James Kerr wrote:
>
> I missed the earlier part of this thread, but if the OP did an upgrade,

Very good point.

> rather than clean install, then there is a known bug:
>
> http://qa.mandriva.com/show_bug.cgi?id=32547


I never upgrade, just keep a spare partition around for new installs.
Old install partition will be used on another new install.
I never run/install cooker packages unless I want to learn/test new releases.

Bit Twister wrote:

> On Sat, 17 Nov 2007 16:33:26 +0000, James Kerr wrote:
>>
>> I missed the earlier part of this thread, but if the OP did an upgrade,
>
> Very good point.
>
>> rather than clean install, then there is a known bug:
>>
>> http://qa.mandriva.com/show_bug.cgi?id=32547
>
>
> I never upgrade, just keep a spare partition around for new installs.
> Old install partition will be used on another new install.
> I never run/install cooker packages unless I want to learn/test new
> releases.

I did an upgrade, it didn't work, so I did a clean reinstall. Didn't use a
clean /var from memory, although I backed up my databases and www partition
first. Perhaps something left over? The advice never to run cooker
packages on one's everyday system is good. I do have enough leftover space
to use alternate /var partitions. RAM is less than 1 y.o., but I agree
that a test may be needed.

I have a spare unused 11G partition at the moment. My feeling is to format
it, then do a reinstall using it as /var. My current /var is 49G with only
4.7G used. If I still get problems, I can then run hardware tests.

Doug.

Bit Twister wrote:

> On Sat, 17 Nov 2007 18:28:45 +1100, Doug Laidlaw wrote:
>>
>> Tried that last night (too late at night.) Forgot to make the script
>> executable. Â*Oh well, tonight is another night....
>
> You must be good at programming. I'd like to see that script.
> Having it run in daily/weekly/monthly which may be running in parallel
> is not a trivial task.
>
I just put in cron.daily a script with a shutdown command after a comment.
The idea was to make my cron job work at a time set by the sequence, not by
the clock, but it hasn't shut the computer down once so far.

Doug.

I was just thinking how wonderful life was, when Doug Laidlaw
opened his gob and said:

>Running Mdv 2008. I am getting Rpm database corruption.

I don't know if this is the same problem, but when I try to run the
update software (or anything in the software management section of
configure your computer) I get a message that the database is locked
by another app - it isn't.

--
Cheers,

Guy

** Stress - the condition brought about by having to
** resist the temptation to beat the living daylights
** out of someone who richly deserves it.

On Tue, 20 Nov 2007 21:27:41 +0000, Dr Teeth wrote:
> I was just thinking how wonderful life was, when Doug Laidlaw
> opened his gob and said:
>
>>Running Mdv 2008. I am getting Rpm database corruption.
>
> I don't know if this is the same problem, but when I try to run the
> update software (or anything in the software management section of
> configure your computer) I get a message that the database is locked
> by another app - it isn't.

Yes, have seen that when Mandriva Online (mdkonline) is checking for updates.
I have not located where mdkapplet is running from.
doing a
grep mdkapplet /var/log/messages
shows time stamps which vary. Wondering if they are 24 hours since the
last urpmi run.
It have not been enough trouble to me to read /usr/bin/mdkapplet

On Tue, 20 Nov 2007 16:45:40 -0500, Bit Twister wrote:

> I have not located where mdkapplet is running from.

/usr/share/autostart/mandriva-mdvonline.desktop

Setting AUTOSTART to FALSE in
~/.MdkOnline/mdkonline will disable it. This is what is done, if you
right click on it, and disable it that way.

Regards, Dave Hodgins

--
Change nomail.afraid.org to ody.ca to reply by email.
(nomail.afraid.org has been set up specifically for
use in usenet. Feel free to use it yourself.)

Doug Laidlaw wrote:

> Bit Twister wrote:
>
>> On Sat, 17 Nov 2007 16:33:26 +0000, James Kerr wrote:
>>>
>>> I missed the earlier part of this thread, but if the OP did an upgrade,
>>
>> Very good point.
>>
>>> rather than clean install, then there is a known bug:
>>>
>>> http://qa.mandriva.com/show_bug.cgi?id=32547
>>
>>
>> I never upgrade, just keep a spare partition around for new installs.
>> Old install partition will be used on another new install.
>> I never run/install cooker packages unless I want to learn/test new
>> releases.
>
> I did an upgrade, it didn't work, so I did a clean reinstall. Didn't use
> a clean /var from memory, although I backed up my databases and www
> partition
> first. Perhaps something left over? The advice never to run cooker
> packages on one's everyday system is good. I do have enough leftover
> space
> to use alternate /var partitions. RAM is less than 1 y.o., but I agree
> that a test may be needed.
>
> I have a spare unused 11G partition at the moment. My feeling is to
> format
> it, then do a reinstall using it as /var. My current /var is 49G with
> only
> 4.7G used. If I still get problems, I can then run hardware tests.
>
> Doug.

No problems with new /var, except as noted elsewhere, and a lot of little
details that re-using the old /var normally took care of. Perhaps Jim was
right. Must do a system backup so that I can simply put everything back.

The absence of default Php (Phpgedview requires a very complete version)
leaves me guessing a bit and wondering how people with servers get on.
Mandriva doesn't seem to have an "enterprise" version, or a server version
like Ubuntu.

Doug.