Text 647, 68 rader
Skriven 2007-03-28 10:33:17 av mark lewis (1:3634/12.0)
    Kommentar till text 646 av Peter Knapper (3:772/1.10)
Ärende: making backups
======================
 PK> If I remember correctly, the catch with the Warp3 32 bit TCP/IP
 PK> upgrade is that it only affects the stack, and not the TCP/IP
 PK> utils, although this is not usually a problem if you use 3rd party
 PK> utils.

 ml> hummm... what utils?

 PK> The TCP/IP Applications like FTP, SMTP, etc...

oh... i don't use the stuff from IBM... never have, really... peter moylan's
ftpser and weasel, brian harvard's apache stuff, a emx compiled version of the
BSD inetd, time868...

now, things like ping, netstat, tracerte, etc... they're all the default
stuff... i guess those are the tools or utils?

 ml> right now, i'm stuck about halfway thru the updates... 
 ml> i don't know if it has made it to the 32bit stack yet 
 ml> or not but it is up and running... kinda... i'm not 
 ml> going to be firing up all my services until i get all 
 ml> the way thru it, though...

 PK> What version of OS/2 are you runing (+FP Level?)

well, when i started, it was Warp 3 Connect w/FP40 and nothing else... no
device driver updates, no tcp/ip updates, and no mpts updates...

 PK> On the comandline, run INETVER and post the output.

it was the default 16bit stack that came with Warp 3 Connect...

however, i'm glad to say that the updates have finally been worked thru and
completed... along the way, we discovered a few bugs that were squashed... now
that box is running the 32bit stack with the necessary mpts update and i
believe there was also another FP that was applied ;)

i'm also glad to say that it hasn't been knocked down in the 12 hours it has
been since it was completed... i was also able to refine my bootup sequence
such that my netinit is initiated from startup.cmd and i've another .cmd that
waits in a holding pattern until the mapped network drives are available before
firing up everything else... this has also afforded me the relief of not having
to be present during the reboot to kick apache and all the nodes of the bbs
into operation... now, apache fires up as one of the first apps, then the rest
of the basic operation and services (in alphabetical order O:) ) and finally
the tradewars bot servers for the four additional alien races and lastly all of
the nodes of the bbs...

oh yeah... there was a time, in the last days i've been working on these
updates, that i didn't have apache running and got the same trap that i had
been seeing... said it was a trap d, indicated it was a trap e, and pointed to
the same place by CSLIM... what i did have running was the weasel smtp server,
the inetd for the gigo smtp server (forwarded to from weasel), and the binkd
server... on the reboot, chkdsk ran and recovered weasel.$$$ which is a temp
logfile... i went and looked and saw where some "script kiddie" was trying a
dictionary attack on weasel... evidently trying to find an account to log in
with so as to be able to send a bunch of spam... i don't know that for sure as
they may have been probing for a hole thru weasel to the shell... i suspect
that whatever it was they were trying to do was causing weasel (if it was
weasel) to make a call to the 16bit stack that it couldn't handle...

it wasn't the inetd or gigo's smtp because inetd shows me the output of that
smtp when it runs... i also doubt that it was the binkd server but it is
possible... all i have/had to go on was the trap that was traced to AFINET.SYS
which has now been replaced with a much newer version O:)

)\/(ark

 * Origin:  (1:3634/12)