valid HOSTNAME characters in syslog?

General discussions here

Moderator: rgerhards

valid HOSTNAME characters in syslog?

Postby rsys on Wed Aug 08, 2007 8:39 pm

RFC 3164 refers to another RFC for the validity of the HOSTNAME field in syslog message. I couldn't find the proper info in the referred RFC. What characters does rsyslog treats as valid HOSTNAME field in syslog message to be relayed?

In particular, Is '@' a valid character in HOSTNAME field of syslog message? SyslogNg treats '@' a valid HOSTNAME character. Not sure what is their source of information.

appreciate any pointers.
rsys
New
 
Posts: 2
Joined: Wed Aug 08, 2007 8:32 pm

Professional Services Information

  • Custom written rsyslog.conf?
  • Maintenance Contract?
  • Installation support?

RE: valid HOSTNAME characters in syslog?

Postby rgerhards on Wed Aug 08, 2007 9:51 pm

I need to check, but we are quite liberal in what we accept. Technically, @ is an invalid character, but I think it's still allowed. Will post when I know exactly.
User avatar
rgerhards
Site Admin
 
Posts: 1780
Joined: Thu Feb 13, 2003 11:57 am

RE: valid HOSTNAME characters in syslog?

Postby rgerhards on Wed Aug 08, 2007 10:00 pm

OK, I checked. '@' is valid, as are a lot of other characters, which are not permitted in hostnames. You may have a look at the code here:

http://rsyslog.cvs.sourceforge.net/rsys ... kup#l_3016

Hostname is delimited by ' ', ':'. These identify a hostname to acutally be a tag: '[', ']' and '/'.

HTH,
Rainer
User avatar
rgerhards
Site Admin
 
Posts: 1780
Joined: Thu Feb 13, 2003 11:57 am

RE: valid HOSTNAME characters in syslog?

Postby rsys on Wed Aug 08, 2007 11:12 pm

Thanks for your investigation. Do you have any reference to RFC which states '@' is indeed a valid character allowed in HOSTNAME field of syslog?
rsys
New
 
Posts: 2
Joined: Wed Aug 08, 2007 8:32 pm

RE: valid HOSTNAME characters in syslog?

Postby rgerhards on Thu Aug 09, 2007 8:22 am

It is *not* a valid character, we just accept it. Please note that RFC 3164 is only an informational document, not a standard. Also, if you read it closely, it simply allows anything. RFC 3195 (a real standard) has made some restrictions. 3195 is not yet of any relevance in practice. There is a new standardization effort ongoing, which's outcome is quite in doubt. See the IETF syslog-sec home page if you'd like to further investigate.

Also, if you provide me some more information on why this question is important to you, I can probably provide some better advise.

Thanks,
Rainer
User avatar
rgerhards
Site Admin
 
Posts: 1780
Joined: Thu Feb 13, 2003 11:57 am

Google Ads



Return to General

Who is online

Users browsing this forum: No registered users and 1 guest

cron