ntp issue invalid group [Solved]

[WeirDave]

* Starting ntpd...
chown: `ntp:ntp': invalid group

Not sure but when ntp starts I get the above error message
_________________
\\eir]>ave
http://www.weirdave.com &
http://aatrade.weirdave.com
Home of The Library and The eXpanse

[speed_bump]

Check the file /etc/group and see if there's a line for ntp. If not, that's your problem. It should be sufficient to add a group for ntp to eliminate the problem.

[WeirDave]

[pjp]

Moved from Other Things Gentoo.
_________________
Quis separabit? Quo animo?

[speed_bump]

No problem. It looks as though this isn't actually stopping ntp from working, just generating an error message. However, to fix it you can add a line to /etc/group. It may differ from this, but on my gentoo box the entry is:

ntp:x:123:

field 1 - Group name
field 2 - Group password (seldom used, x is fine)
field 3 - Group ID

If you look at /etc/passwd, you should see a line for ntp. The quickest way to check is:

grep ntp /etc/passwd

you should get something that looks like this:

ntp:x:123:123:added by portage for ntp:/dev/null:/bin/false

Here's a breakdown. Forgive me if this is too basic, but I'd rather provide too much information rather than too little

field 1 - Username
field 2 - password - x indicates that shadow passwords are in use
field 3 - User ID - the numeric value the system uses to uniquely identify users
field 4 - Group ID - the numeric value the system uses to identify groups
field 5 - GECOS - a free form string that typically describes the account
field 6 - Home directory - in this case, /dev/null since we never expect this user to log in
field 7 - login shell - in this case, /bin/falsel since we never expect this user to log in

Your best bet is to make the Group ID field in /etc/group the same as the Group ID field in /etc/passwd.

[WeirDave]

[speed_bump]

If you have ntpd running (ps -ef | grep -i ntp) then it should continuously udate the time as long as it's configured and working directly. However, if the clock is wildly inaccurate (not uncommon if the machine has been powered off for a while, or has not used NTP previously), ntpd will not adjust the clock. It will write a message to syslog and exit. In addition, if the clock is way off, but within the panic threshold which is 1000 seconds, it will slowly adjust the time toward the correct time. The man page for ntpd indicates this could take up to 14 days (!) for a 600 second skew.

To avoid this, you can either use

ntpdate (deprecated)

or

ntpd -q

to get the clock to the right time, and then ntpd will keep it accurate after that. Frequently, the ntpd -q will be run in a startup script to ensure a good start point. The only gotcha with all of that is if you're using things that are highly date sensitive (using make in NFS mounted directories would be the most common example), wild time changes could lead to "interesting" results.

hth