atd and load question

atd and load question

Post by shau.. » Sat, 14 Sep 2002 07:16:14



okay, got a linux question about 'atd'.

i'm running RedHat Linux 7.2 kernel 2.4.7-10
on an intel box (dell for what it's worth).

i have a job that creates a shell script
and batches it.  the script goes into
the /var/spool/at directory (which, by the
way, is NOT NFS mounted).

the linux server is always busy and never
seems to run the job.  i look and the man
page for atd says, 'atd will run the job by
default if the loadavg is about .8;  use
the '-l' option to make adjustments'.  i look
at /proc/loadavg and it hangs around 3.xx ...

so ... i configure the atd script to
run with a pre-defined load average
of 4 ("atd -l 4" ... our system rarely goes
over that).

now, when a job is submitted and is
placed in the /var/spool/at directory,
the script disappears (i assume that it's being
processed) ... BUT ... when i search for the
job (i have the job number) or 'ps auwwx'
around, i can find no trace of the job, the
job number, the process, or anything related.
even when i type 'atq', i get nothing.

other items of interest:

-- when the job was placed in the /var/spool/at
directory, the script is owned by 'nobody' and
the group is some wild 7 digit number.  i didn't
think that would matter, but it may.

-- also, the permissions and owner for the /var/spool,
/var/spool/at and /var/spool/at/spool are:

drwxr-xr-x   14 root     root         4096 Jan 29  2002 spool

drwx------    3 daemon   daemon       4096 Sep 12 17:23 at

drwx------    2 daemon   daemon       4096 Sep 12 17:23 spool

-- there doesn't seem to be ANY errors in /var/log/messages

-- lastly, when i launch at job by hand ('at now blah, blah, blah')
the job is put in queue and runs okay.

any suggestions?

-X

 
 
 

atd and load question

Post by Duane Smecker » Fri, 20 Sep 2002 07:10:40


would the wild 7 didit number be something like 67198863?
That would be -1 in some systems.  It probably means the
group was not valid for whatever created the script.

Next, the shell script should run the batch script with the
at command running in batch mode.
with something like  
    at -f /tmp/scriptname teatime
(which runs the script called scriptname at 4pm)

MAKE SURE ITS AN EXECUTABLE SCRIPT FOR NOBODY!


>okay, got a linux question about 'atd'.

>i'm running RedHat Linux 7.2 kernel 2.4.7-10
>on an intel box (dell for what it's worth).

>i have a job that creates a shell script
>and batches it.  the script goes into
>the /var/spool/at directory (which, by the
>way, is NOT NFS mounted).

>the linux server is always busy and never
>seems to run the job.  i look and the man
>page for atd says, 'atd will run the job by
>default if the loadavg is about .8;  use
>the '-l' option to make adjustments'.  i look
>at /proc/loadavg and it hangs around 3.xx ...

>so ... i configure the atd script to
>run with a pre-defined load average
>of 4 ("atd -l 4" ... our system rarely goes
>over that).

>now, when a job is submitted and is
>placed in the /var/spool/at directory,
>the script disappears (i assume that it's being
>processed) ... BUT ... when i search for the
>job (i have the job number) or 'ps auwwx'
>around, i can find no trace of the job, the
>job number, the process, or anything related.
>even when i type 'atq', i get nothing.

>other items of interest:

>-- when the job was placed in the /var/spool/at
>directory, the script is owned by 'nobody' and
>the group is some wild 7 digit number.  i didn't
>think that would matter, but it may.

>-- also, the permissions and owner for the /var/spool,
>/var/spool/at and /var/spool/at/spool are:

>drwxr-xr-x   14 root     root         4096 Jan 29  2002 spool

>drwx------    3 daemon   daemon       4096 Sep 12 17:23 at

>drwx------    2 daemon   daemon       4096 Sep 12 17:23 spool

>-- there doesn't seem to be ANY errors in /var/log/messages

>-- lastly, when i launch at job by hand ('at now blah, blah, blah')
>the job is put in queue and runs okay.

>any suggestions?

>-X

--
Duane Smeckert
google for email
spambots, you should not send email



 
 
 

1. atd problem: empty atd.pid

When starting Linux, the message "Starting at services.." appears with
"done". But the file /var/run/atd.pid remains empty. This causes
(probably) that remote printing from other machines is not possible.
When I shutdown, an error message (empty file /var/run/atd.pid) occurs.

The problem occurs since 1 month. Who can help me?
--
Klaus Bartke

2. Severe SCSI-error

3. failed loading of atd and /etc/rc.d/rc.local

4. Install 2.6 almost without downtime?

5. Matrox Mystique ands X.

6. Last process stored in $!

7. atd daemon

8. Redhat 6.0 says it supports s3trio3d BUT

9. atd: Redirecting stdout failed. Permission denied.

10. atd daemon glitch?

11. What good are portmap, routed, and atd?

12. Are apmd and atd necessary?

13. upgrade resistant atd,chsh2,top,netstat,...