Re: 20010126: LDM logfiles and server access

Erick Lorenz wrote:
> 
> 
> This is the line that I have:
> local0.err;local0.warning;local0.notice;local0.info
> /usr/local/ldm/logs/ldmd.log
> 

Hi Eric,

This will work, although you won't see messages at the debug level.  If
you ever need to run the ldm in debug mode, you'd want to see those
messages.

(Btw, I assume that syslog on Irix is similar to the common BSD syslog
behavior where all messages of the specified priority and higher are
logged according to the given action, and thus, you will indeed get
messages of level CRIT or higher.  If this is the case, you really only
need "local0.info" in the entry - the others are redundant.  I'd advise
you to change it to simply "local0.debug".)


> 
> After checking all these points I tried sending a kill -HUP to syslogd.
> After this
> I started the ldm and lo, the log file started to fill up.  Now I can see
> that I am
> still getting access denied messages
> 

This raises the question of whether or not the ldm  is able to HUP the
syslog daemon with its 'hupsyslog' program.  

In looking at the hupsyslog code I see a special bit of code for an
SGI.  The comments say something regarding the file /etc/syslog.pid not
existing on that system.  Is that true for you?  If so, then maybe
there's a problem with this bit of code...  although logging used to
work for you, right?

You could test whether the ldm can hup the syslog daemon.  Let me know
if you want ideas about how to do this.


> Jan 27 01:29:27 5Q:atm23 rpc.ldmd[1196]: Starting Up (built: Jan 19 1999
> 15:57:11)
> Jan 27 01:29:27 5Q:atm23 pqexpire[1197]: Starting Up
> Jan 27 01:29:27 5Q:atm23 typhoon[1201]: run_requester: Starting Up:
> typhoon.atmos.ucla.edu
> Jan 27 01:29:27 5Q:atm23 typhoon[1201]: run_requester: 20010127002927.717
> TS_ENDT {{DDPLUS,  ".*"}}
> Jan 27 01:29:27 5Q:atm23 pqbinstats[1198]: Starting Up (1196)
> Jan 27 01:29:27 5Q:atm23 pqact[1199]: Starting Up
> Jan 27 01:29:27 3Q:atm23 typhoon[1201]: FEEDME(typhoon.atmos.ucla.edu): 7:
> Access denied by remote server
> Jan 27 01:29:27 5Q:atm23 pqsurf[1200]: Starting Up (1196)
> Jan 27 01:29:27 5Q:atm23 pqact[1203]: Starting Up
> Jan 27 01:29:28 5Q:atm23 udp.ldmd[1202]: Starting Up
> Jan 27 01:29:29 5Q:atm23 localhost[1207]: Connection from localhost
> Jan 27 01:29:29 5Q:atm23 localhost[1207]: Connection reset by peer
> Jan 27 01:29:29 5Q:atm23 localhost[1207]: Exiting
> Jan 27 01:29:57 5Q:atm23 typhoon[1201]: run_requester: 20010127002957.806
> TS_ENDT {{DDPLUS,  ".*"}}
> Jan 27 01:29:57 3Q:atm23 typhoon[1201]: FEEDME(typhoon.atmos.ucla.edu): 7:
> Access denied by remote server
> 
> etc,etc
> 
> ====================================
> >Regarding not receiving data, the command
> >       notifyme -vl - -h typhoon.atmost.ucla.edu
> >will show you two things.  If typhoon has an 'allow' line for your machine
> >in it's configuration, which is what you need so that it will relay data
> >to you, then it will show you all the data that is arriving there.
> 
> Here is what I get:
> ldm> notifyme -vl - -h typhoon.atmos.ucla.edu
> Jan 27 01:40:36 notifyme[1261]: Starting Up: typhoon.atmos.ucla.edu:
> 20010127014036.626 TS_ENDT {{ANY,  ".*"}}
> Jan 27 01:40:36 notifyme[1261]: NOTIFYME(typhoon.atmos.ucla.edu): 7: Access
> denied by remote server
> Jan 27 01:41:01 notifyme[1261]: NOTIFYME(typhoon.atmos.ucla.edu): 7: Access
> denied by remote server
> 
> snd it keeps on reporting.
> 
> Is the message actually sent by the remote server or is it generated by the
> local notify me program based on its ability to communicate with the
> server.  Since James says that he has my computer in his allow file I
> wonder if I could be blocked at some other point.  I can ping typhoon.
> 

The answer is 'b'.  I see I have received another email from you
regarding your upstream feed.  I'll respond to that one.


> >If your syslog daemon got bolluxed up, and something else changed causing
> >you to stop receiving data, it sounds possible that you might be having
> >some system problems.  Has anyone checked the system log? (usually in
> >/var/adm/messages)
> 
> I can't see anything in my message file that sticks out except that earlier
> I was having some write errors to the disk that takes the ldm data but I
> don't see how that would affect my access.  Earlier I had that problem and
> the ldm kept ingesting without a hitch.
> 

Yes, I wouldn't think a disk write failure would affect access to your
upstream host.


> Thanks
> 
> Erick
> 

You're welcome!

Anne
-- 
***************************************************
Anne Wilson                     UCAR Unidata Program            
address@hidden                 P.O. Box 3000
                                  Boulder, CO  80307
----------------------------------------------------
Unidata WWW server       http://www.unidata.ucar.edu/
****************************************************