Hobbit Client for SCO 🔗 link

Anyone out there have a compiled SCO client?  Im looking for a version
that would run on 5.0.4, 5.0.5, 5.0.6 

Thanks

Trent

Trent Melcher wrote :

Hi,

I have one, built on 5.0.5 iirc.

It is available on : http://charles.goyard.free.fr/hobbit/, along with a
small install script.

-- 
Charles Goyard - user-a6cdca7046e2@xymon.invalid - (+33) 1 45 38 01 31
Orange Business Services - online multimedia  // ingénierie

Hi All,

More of a curiosity than anything else but has anyone come up with any
sort of xxmbps/client for an average amount of network traffic from a
client to server as well as a average for a proxy to another server
(ours uses between 2-10mbps (usually 2) for approx 150 clients).

Reason I'm asking is to show off hobbit to the phbs so just wondering if
there was any "hard numbers"

Thanks,
Jason.

On Thu, 2007-08-23 at 10:14 +0200, Charles Goyard wrote:

Thanks, installed and working.  Did you make any modifications to get
the memory to report correctly?

My memory isnt reporting correctly, here is the info sent from the
client to the hobbit server.

So Im guessing something on the Hobbit server side isnt set right for
SCO

[memsize]
[freemem]
Average      22814    860000     28216    130372
[swap]
path               dev  swaplo blocks   free
/dev/swap          1,41      0 860000 860000


Thanks again for the client
Trent

It is available on : http://charles.goyard.free.fr/hobbit/, along with a
small install script.

Trent Melcher wrote :

It seems you lack /etc/memsize. I just put it for you on the same web
site (http://charles.goyard.free.fr/hobbit/)

Here are the modes :
-r-xr-xr-x   1 root     bin 		memsize


I'm somehow surprised someone else still has SCO :)

-- 
Charles Goyard - user-a6cdca7046e2@xymon.invalid - (+33) 1 45 38 01 31
Orange Business Services - online multimedia  // ingénierie

On Thu, 2007-08-23 at 17:41 +0200, Charles Goyard wrote:

Thanks

I'm somehow surprised someone else still has SCO :)

Its old technology but it still works and does what we need it to, we
use it for our Dialogic and Brooktrout cards...the code that manages the
cards for dialing and recordings was written for SCO and nobody wants to
spend the time and try and port it over to Linux.

Trent

I got asked similar question about hobbit client.
Can we come up with a metric tool to measure how costly  hb client using  
resources ?

T.J. Yang

See what you�re getting into�before you go there 
http://newlivehotmail.com/?ocid=TXT_TAGHM_migration_HM_viral_preview_0507

On Thu, Aug 23, 2007 at 09:30:38AM +0100, Jones, Jason (Altrincham) wrote:

Hobbit client -> Hobbit server: It sends the ~hobbit/client/tmp/msg*.txt
file once every 5 minutes. Typically this is something like 50 kB which 
amounts to 50*1024*8 / 300 = 1.4 kbit/sec

For the proxy, look at the "bbproxy" status column. It has counts for
the number of messages in and out; it doesn't provide byte counts, but
normal status messages are usually a few kB each. If it's a dedicated
server that handles the proxying, you can just look at the network
utilisation graph.


Regards,
Henrik

On Thu, Aug 23, 2007 at 02:17:34PM -0500, T.J. Yang wrote:

On a completely idle system, run vmstat with 1 second display updates
while the client is active. But my guess is that you won't see anything
at all - the client really uses very few ressources (it's only a few
simple OS commands).


Regards,
Henrik

Thanks again for the SCO client,  its been working great....I do have
one question, did you have to modify anything for the TCP/IP Statistics
on the TRENDS page? On mine it look as if the Retrans and the Packets
Out are are switched.  Looking through the code I cant find the exact
reference where its identifying the variable to the proper output line.

Thanks
Trent

On Thu, 2007-08-23 at 17:41 +0200, Charles Goyard wrote:

Ignore my last post,  I think I found it.

In hobbitd/rrd/do_netstat.c

I think these 2 should be reversed -  see below
        "^[\t ]*([0-9]+) data packets \\(([0-9]+) bytes\\)$",
        "^[\t ]*([0-9]+) data packets \\(([0-9]+) bytes\\) retransmitted

That way they match the correct pickdata subroutine.


do_netstat.c lines 88-95

              case AT_TCP:
  if (pickdata(datapart, pcreset[0],  0, &tcpretranspackets,
&tcpretransbytes)||
      pickdata(datapart, pcreset[1],  0, &tcpoutdatapackets,
&tcpoutdatabytes)   ||
      pickdata(datapart, pcreset[2],  0, &tcpinorderpackets,
&tcpinorderbytes)   ||
      pickdata(datapart, pcreset[3],  0, &tcpoutorderpackets,
&tcpoutorderbytes) ||
      pickdata(datapart, pcreset[4],  0, &tcpconnrequests) ||
      pickdata(datapart, pcreset[5],  0, &tcpconnaccepts)) havedata++;
                                break;

do_netstat.c lines 140-154
static const char *netstat_sco_sv_exprs[] = {
        /* TCP patterns */
        "^[\t ]*([0-9]+) data packets \\(([0-9]+) bytes\\)$",
        "^[\t ]*([0-9]+) data packets \\(([0-9]+) bytes\\) retransmitted
$",
        "^[\t ]*([0-9]+) packets \\(([0-9]+) bytes\\) received
in-sequence$",
        "^[\t ]*([0-9]+) out-of-order packets \\(([0-9]+) bytes\\)$",
        "^[\t ]*([0-9]+) connection requests$",
        "^[\t ]*([0-9]+) connection accepts$",
        /* UDP patterns */
        "^[\t ]*([0-9]+) incomplete headers$",
        "^[\t ]*([0-9]+) bad data length fields$",
        "^[\t ]*([0-9]+) bad checksums$"
        "^[\t ]*([0-9]+) input packets delivered$",
        "^[\t ]*([0-9]+) packets sent$"

};

Trent

On Thu, 2007-08-23 at 17:41 +0200, Charles Goyard wrote: