[SOLVED] ntp crontab duplicated

businesscat · 06-27-2017, 03:37 AM

Hi everyone,
I'm configuring a crontab in my new host and it shows the following messages on the /var/log/cron:

Code:

Jun 27 06:07:01 SERVER CROND[12059]: (root) CMD (root /usr/sbin/ntpdate ntp0 > /dev/null 2>&1)
Jun 27 06:07:01 SERVER CROND[12060]: (root) CMD (/usr/sbin/ntpdate ntp0 > /dev/null 2>&1)
Jun 27 07:01:01 SERVER CROND[12641]: (root) CMD (run-parts /etc/cron.hourly)
Jun 27 07:01:01 SERVER run-parts(/etc/cron.hourly)[12641]: starting 0anacron
Jun 27 07:01:01 SERVER run-parts(/etc/cron.hourly)[12650]: finished 0anacron

as you can see, is executing two times...and I don´t know the reason...

my cron file is the following:

Code:

[root@SERVER cron.d]# cat /etc/cron.d/
0hourly     ntp         raid-check  
[root@SERVER cron.d]# cat /etc/cron.d/ntp 
#sincronize hour 
#Linux system
7 * * * * root /usr/sbin/ntpdate ntp0 > /dev/null 2>&1

OS:
Red Hat Enterprise Linux Server release 6.9 (Santiago)

Permisions on the file:

Code:

-rw-r--r--   1 root root   93 Jun  6 13:58 ntp

I check other hosts, not configuring by me, and show the following log:

Code:

Jun 27 05:00:01 SERVER crond[823]: (*system*) NULL security context for user, but SELinux in permissive mode, continuing ()
Jun 27 05:00:01 SERVER crond[824]: (root) CMD (/usr/sbin/ntpdate ntp0 >/dev/null)
Jun 27 05:01:01 SERVER crond[1713]: (*system*) NULL security context for user, but SELinux in permissive mode, continuing ()
Jun 27 05:01:01 SERVER crond[1714]: (root) CMD (run-parts /etc/cron.hourly)

Thank you in advance

scasey · 06-27-2017, 05:37 AM

Quote:

Originally Posted by businesscat

Hi everyone,
I'm configuring a crontab in my new host and it shows the following messages on the /var/log/cron:

Code:

Jun 27 06:07:01 SERVER CROND[12059]: (root) CMD (root /usr/sbin/ntpdate ntp0 > /dev/null 2>&1)
Jun 27 06:07:01 SERVER CROND[12060]: (root) CMD (/usr/sbin/ntpdate ntp0 > /dev/null 2>&1)
Jun 27 07:01:01 SERVER CROND[12641]: (root) CMD (run-parts /etc/cron.hourly)
Jun 27 07:01:01 SERVER run-parts(/etc/cron.hourly)[12641]: starting 0anacron
Jun 27 07:01:01 SERVER run-parts(/etc/cron.hourly)[12650]: finished 0anacron

as you can see, is executing two times...and I don´t know the reason...

my cron file is the following:

Code:

[root@SERVER cron.d]# cat /etc/cron.d/
0hourly     ntp         raid-check  
[root@SERVER cron.d]# cat /etc/cron.d/ntp 
#sincronize hour 
#Linux system
7 * * * * root /usr/sbin/ntpdate ntp0 > /dev/null 2>&1

[snip]
Thank you in advance

What is in /etc/cron.d/0hourly ?

businesscat · 06-27-2017, 07:07 AM

Quote:

Originally Posted by scasey

What is in /etc/cron.d/0hourly ?

this:

SHELL=/bin/bash
PATH=/sbin:/bin:/usr/sbin:/usr/bin
MAILTO=root
HOME=/
01 * * * * root run-parts /etc/cron.hourly

michaelk · 06-27-2017, 07:20 AM

It looks like you have two separate cron jobs. Check roots cron jobs i.e.
crontab -l

businesscat · 06-27-2017, 07:29 AM

Quote:

Originally Posted by michaelk

It looks like you have two separate cron jobs. Check roots cron jobs i.e.
crontab -l

Code:

[root@HOST~]# crontab -l
# Sincro 
# Linux
7 * * * * root /usr/sbin/ntpdate ntp0 > /dev/null 2>&1

only appear this cron

michaelk · 06-27-2017, 07:41 AM

You have two separate cron jobs running. One as the root user and one as a system job i.e. /etc/cron.d.

Delete the one running under root. You can use crontab -e.

In addition why are you running ntpdate versus ntp?

businesscat · 06-27-2017, 07:59 AM

Quote:

Originally Posted by michaelk

You have two separate cron jobs running. One as the root user and one as a system job i.e. /etc/cron.d.

Delete the one running under root. You can use crontab -e.

In addition why are you running ntpdate versus ntp?

I see, but one question:

How can I see where is the other crontab is defined?

If i use crontab -e appears the following:

Code:

# Sincro
# Linux
7 * * * * root /usr/sbin/ntpdate ntp0 > /dev/null 2>&1

If I comment the last line, I suppose it doesn't apply anymore...right? But...where is the other crontab defined?

More info:

I create this crontab with the following commands:

Code:

vim /etc/cron.d/ntp
 
contrab ntp

thank you so much, probably is a stupid thing but I can´t see that...too much hours with Windows...

businesscat · 06-27-2017, 08:17 AM

Quote:

Originally Posted by michaelk

In addition why are you running ntpdate versus ntp?

When I execute this ntpdate:

Code:

[root@HOSTcron.d]# /usr/sbin/ntpdate ntp0
27 Jun 14:16:16 ntpdate[48823]: adjust time server 10.59.172.47 offset -0.031481 sec

michaelk · 06-27-2017, 08:18 AM

One cron job is /etc/cron.d/ntp. This is a system cron job

Second cron job is under root user (crontab -l). This is a users cron job which are saved to /var/spool/cron by default.

ntpdate will force a time jump while ntp slowly adjusts time.

businesscat · 06-27-2017, 08:46 AM

Ok, I´m (think) understand your point. Tomorrow I will test your advice.

Thank you so much

for your patience.

scasey · 06-27-2017, 11:38 AM

Quote:

Originally Posted by businesscat

this:

SHELL=/bin/bash
PATH=/sbin:/bin:/usr/sbin:/usr/bin
MAILTO=root
HOME=/
01 * * * * root run-parts /etc/cron.hourly

What is in /etc/cron.hourly?

michaelk · 06-27-2017, 12:07 PM

Although we do not have all of the cron log entries and making a few assumptions but if the OP put their ntp script in cron.hourly there would be a log entry like.

Quote:

Jun 27 07:01:01 SERVER run-parts(/etc/cron.hourly)[12641]: starting ntp
Jun 27 07:01:01 SERVER run-parts(/etc/cron.hourly)[12650]: finished ntp

Quote:

Jun 27 06:07:01 SERVER CROND[12059]: (root) CMD (root /usr/sbin/ntpdate ntp0 > /dev/null 2>&1)

This would indicate a cron.d or a users cron job.

Quote:

(root /usr/sbin/ntpdate ntp0 > /dev/null 2>&1)

This is the command that cron will run. System cron jobs require a run as user but a regular user's does not.
Since root shows up in the command it typically means that there is a root cron job.

scasey · 06-27-2017, 01:53 PM

Quote:

Originally Posted by michaelk

This is the command that cron will run. System cron jobs require a run as user but a regular user's does not.
Since root shows up in the command it typically means that there is a root cron job.

Ahh. I wondered what the difference was between these:

Code:

Jun 27 06:07:01 SERVER CROND[12059]: (root) CMD (root /usr/sbin/ntpdate ntp0 > /dev/null 2>&1)
Jun 27 06:07:01 SERVER CROND[12060]: (root) CMD (/usr/sbin/ntpdate ntp0 > /dev/null 2>&1)

So the first indicates the user root's crontab, and the second is from /etc/cron.d/ntp -- right?
The OP is apparently logged in as root, so when he does

Code:

crontab -l

he is seeing the second cron, yes?
Then he should probably just

Code:

crontab -e

and delete that one. Agreed?

michaelk · 06-27-2017, 02:09 PM

Yes, however if the root cron job did not have "root" in the command then the log would look the same as the /etc/cron.d

businesscat · 06-28-2017, 02:53 AM

Quote:

Originally Posted by scasey

What is in /etc/cron.hourly?

This:

Code:

[root@HOST~]# cat /etc/cron.hourly/0anacron 
#!/bin/bash
# Skip excecution unless the date has changed from the previous run 
if test -r /var/spool/anacron/cron.daily; then
    day=`cat /var/spool/anacron/cron.daily`
fi
if [ `date +%Y%m%d` = "$day" ]; then
    exit 0;
fi

# Skip excecution unless AC powered
if test -x /usr/bin/on_ac_power; then
    /usr/bin/on_ac_power &> /dev/null
    if test $? -eq 1; then
    exit 0
    fi
fi
/usr/sbin/anacron -s

Quote:

Originally Posted by scasey

Ahh. I wondered what the difference was between these:

Code:

Jun 27 06:07:01 SERVER CROND[12059]: (root) CMD (root /usr/sbin/ntpdate ntp0 > /dev/null 2>&1)
Jun 27 06:07:01 SERVER CROND[12060]: (root) CMD (/usr/sbin/ntpdate ntp0 > /dev/null 2>&1)

So the first indicates the user root's crontab, and the second is from /etc/cron.d/ntp -- right?
The OP is apparently logged in as root, so when he does

Code:

crontab -l

he is seeing the second cron, yes?
Then he should probably just

Code:

crontab -e

and delete that one. Agreed?

I comment the line on crontab -e

Code:

#7 * * * * root /usr/sbin/ntpdate ntp0 > /dev/null 2>&1

and, on the log shows:

Code:

Jun 28 08:07:01 HOST CROND[60678]: (root) CMD (/usr/sbin/ntpdate ntp0 > /dev/null 2>&1)
Jun 28 08:47:58 HOST crontab[61109]: (root) BEGIN EDIT (root)
Jun 28 08:48:07 HOST crontab[61109]: (root) END EDIT (root)

Quote:

Originally Posted by michaelk

Yes, however if the root cron job did not have "root" in the command then the log would look the same as the /etc/cron.d

Thank you michalek

In your opinion, what is the best way to create a crontab?

I use the following:

Code:

vim /etc/cron.d/ntp

add to the file:
7 * * * * root /usr/sbin/ntpdate ntp0 > /dev/null 2>&1

 
contrab ntp

Thank you, by the moment the log shows:

Code:

Jun 28 08:07:01 HOST CROND[60678]: (root) CMD (/usr/sbin/ntpdate ntp0 > /dev/null 2>&1)
Jun 28 08:47:58 HOST crontab[61109]: (root) BEGIN EDIT (root)
Jun 28 08:48:07 HOST crontab[61109]: (root) END EDIT (root)