[Beowulf] Regarding wulfstat

Robert G. Brown rgb at phy.duke.edu
Fri May 14 23:41:27 EDT 2004


On Fri, 14 May 2004, Jon Tegner wrote:

> Thanks a lot for info!
> 
> After doing the verbose test I realized that the nodes that use 
> hostrange got the wrong port number. After specifying 7887 it worked as 
> expected.

This was a bug (you shouldn't need to specify a port for iprange or
hostrange) but I >>think<< it is fixed in 0.5.8, posted on the
website(s) now.

At this point I can't get it to either crash or use the wrong port
number.  Hopefully it is all better again.

   rgb

> 
> Thanks again, much apreciated!
> 
> /jon
> 
> 
> Robert G. Brown wrote:
> 
> >On Fri, 14 May 2004, Jon Tegner wrote:
> >
> >  
> >
> >>Hi,
> >>    
> >>
> >
> >I reproduce the iprange behavior here, so that is just a plain old bug
> >and I'll fix it shortly, I hope.
> >
> >The other I'm having a hard time reproducing, although I don't have a
> >lot of RH7.3 systems left to play with.  
> >
> >While I'm trying, check that wulfstat is current (I just now posted
> >0.5.7 to both brahma and personal sites).  It has been substantially
> >rewritten with the most important change being that it now is based on a
> >library shared with both wulflogger and indirectly wulfweb (to
> >facilitate further development of client tools).
> >
> >It also has better debugging.  If you build it from e.g. the tarball for
> >testing purposes, and put your failing wulfhosts specs into e.g.
> >wulfhosts.tst, then running:
> >
> >  ./wulfstat -f wulfhosts.tst -v 1 2> log
> >
> >should BOTH let you see wulfstat AND create a maximally verbose log of
> >what it is doing.  Let it run for maybe one update and quit wulfstat,
> >then look at the log.  It SHOULD tell you just what is failing in a
> >segment near the top like:
> >
> >D_READHOST: Entering hostrange loop
> >Validating hostname = r00, port = 0
> >D_READHOST: Validating host:
> >D_READHOST: Starting name = |r00| hostip = || inetaddress = |33b60398| port = |0|
> >D_READHOST: Looking up ip number for host r00.
> >D_READHOST: Got host_id = 42134b58.
> >D_READHOST: Setting ip_tmp to 192.168.182.50
> >D_READHOST: Reset hostip from 
> >D_READHOST: hostip is now 192.168.182.50
> >D_READHOST: Setting inetaddr from hostip 192.168.182.50
> >D_READHOST: Setting host r00's port = 7887
> >D_READHOST: Cleaned up host r00 (we hope).
> >D_READHOST: Ending name = r00 hostip = 192.168.182.50 inetaddress = 32b6a8c0  port = 7887
> >Validating hostname = r01, port = 0
> >D_READHOST: Validating host:
> >D_READHOST: Starting name = |r01| hostip = || inetaddress = |32b6a8c0| port = |0|
> >D_READHOST: Looking up ip number for host r01.
> >D_READHOST: Got host_id = 42134b58.
> >D_READHOST: Setting ip_tmp to 192.168.182.51
> >D_READHOST: Reset hostip from 
> >D_READHOST: hostip is now 192.168.182.51
> >D_READHOST: Setting inetaddr from hostip 192.168.182.51
> >D_READHOST: Setting host r01's port = 7887
> >D_READHOST: Cleaned up host r01 (we hope).
> >D_READHOST: Ending name = r01 hostip = 192.168.182.51 inetaddress = 33b6a8c0  port = 7887
> >
> >Obviously it works here, and should clearly indicate it if the problem
> >is e.g. something not resolving.
> >
> >  
> >
> >>Any help/hints on this would be appreciated. And Thank You Robert for a 
> >>nice little program!
> >>    
> >>
> >
> >You're welcome and I'm glad you are enjoying it.  You should check out
> >wulflogger and wulfweb (the latter is pretty heavily beta still, but
> >works) if you haven't.  You can see what wulfweb can do (very crudely
> >still -- don't have time to really pretty this up yet) on:
> >
> >  http://www.phy.duke.edu/resources/computing/brahma/wulfweb/wulfweb.html
> >
> >Our local users like this better than wulfstat in a lot of cases,
> >although I don't have any of the other views done yet.  I don't think it
> >is as useful for admins or serious users, though -- I tend to run it at
> >a time granularity of around a minute because updates bollix old/stupid
> >browsers.
> >
> >I'll get back to you when I've fixed iprange.  This is an egregious
> >failure, so it is probably a simple enough bug.
> >
> >   rgb
> >
> >  
> >
> 
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org
> To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
> 

-- 
Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu



_______________________________________________
Beowulf mailing list, Beowulf at beowulf.org
To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf



More information about the Beowulf mailing list