finger trouble

[Simon Page_1]

The finger command is returning "In real life: ???" for a lot (not all) of the NIS users. Any ideas?

[Rick Garland]

Does the NIS passwd file have this information?

[Pete Randall]

I would guess that that information was never supplied. Probably the original password file, from which the NIS database was created had users in it with no GECOS information, paricularly the Real Name.


Pete

[Alan Meyer_4]

What does NIS have for user information in it's database?

[Simon Page_1]

All the info is in the password file. finger used to work a few days ago - it's just stopped working for most users.

[Alan Meyer_4]

Can we assume that all the other finger information displays correctly and only the "In Real Life" field displays as ???

[Rick Garland]

The /usr/bin/chfn command will change the way the name is displayed when 'finger' is used.

Is this command is uise somewhere, or by someone?

[Simon Page_1]

There's no info present when the finger displays "In real life: ???".
Please see attached file for a couple of examples.

[Alan Meyer_4]

The only time I get that type of response is when the user does not exist in the NIS database or the local passwd file.

[Alan Meyer_4]

I also get that response if the NIS/passwd record is incomplete or in the incorrect format

[Simon Page_1]

Hi Alan, yes I agree with your comments, they would be my thoughts too! That's why this one's got me stumped.

[Alan Meyer_4]

what does ypwhich yield?

Can we assume that the users can login successfully?

[Rick Garland]

Can you run the pwck (passwd check) against the file? Any discrepencies show up on output?

[Simon Page_1]

what does ypwhich yield? The name of the NIS master

Can we assume that the users can login successfully? - YES, everything (seemingly!) works fine except for finger

[Simon Page_1]

Hi Rick, you asked "Can you run the pwck (passwd check) against the file? Any discrepencies show up on output?" - As the NIS passwd file is quite big I created a subset file with just the 2 users (i.e the 2 users in the attachment file I sent) - pwck doesn't report any problems!

[Rick Garland]

An idea.

When you run the finger command and get the "In Real Life ???", I can reproduce that by using a non-existent acct. If I remove the GECOS info from a valid acct and run finger, the "In Real Life" field does not even show. By chance is the finger looking at the wrong passwd file? The local passwd file does not have this acct so it shows the ??? while the NIS passwd is OK?

[Alan Meyer_4]

I agree with something along thouse lines Rick, but how does finger look at the local passwd file only?

I tried to remove the "+" from the end and it still resolves NIS information with finger... From a cache maybe?

[Simon Page_1]

Hi Rick, you have a good point there.
How does finger determine where it's information is coming from?

[Rick Garland]

Do the ypmatch command on the account mcarte71. What is output?

Is mcarte71 in the local passwd file or the NIS passwd?

Is the mcastan7 acct in the local passwd file?

[Simon Page_1]

Hi Rick

Do the ypmatch command on the account mcarte71. What is output?

mcarte71:XecnEYx3R/08w:935724:935724:Mark Carter:/dtv7654/h909/u/mcarte71:/bin/csh

Is mcarte71 in the local passwd file or the NIS passwd?
Only in NIS passwd

Is the mcastan7 acct in the local passwd file? NO

[See attached for a typical /etc/passwd file]

[Rick Garland]

The only reference I am able to find about the ??? is on this site

https://engineering.purdue.edu/ECN/Resources/KnowledgeBase/Docs/20020202100807

Question, does the finger command return the same output when you do this on other systems?
What happens if you append?

finger mcastan7@yourdomain.com

[Simon Page_1]

Some more info!
We also have SunOS 5.6/5.8 and Tru64 5.1 machines - the SunOS machines show the problem but the Tru64 machines don't!!
So (I assume) there cannot be anything inherently wrong in the way that the NIS passwd server etc is set up, it must be the way that finger is handling/interpreting the data.

[Rick Garland]

Append the domain names to your finger commands.

[Alan Meyer_4]

What about an imcomplete/corrupt push of the map files? Can you re-push all the map files? does that solve the problem?

[Rick Garland]

I am beginning to think there is a resolution issue. Are all systems able to resolve? Do they resolve with the FQDN or with the short name? Is this pattern consistent across all of the platforms? (HPUX, SUN, Tru64)