Privacy and Security Notice

Archived Messages for LINUX-USERS@cebaf.gov: SLIP woes solved

SLIP woes solved

Michael Tiefenback (tiefen@CEBAF.GOV)
Sat, 03 Jan 1998 23:17:44 -0500 (EST)

Yes, it was a brain fade issue. I had updated one file on my personal
machine at home on 8/23/97 with a new Jlab modem bank PASSWORD. I
probably also updated the actual file I used on both machines, but in the
modem swapover and copying files around, I probably overwrote that one
with an older copy. At least I used cp -a and the timestamps were useful
to distinguish the date of the contents.

Using MASQ through the 486SX, I was getting almost immediate errors of the
following type

kernel: MASQ: failed TCP/UDP checksum

after ssh'ing in from home. This happened to me for the first three or
four sequential connects from home after updating the password. After
tiring if this, I added an mtu query in the process. I find that
apparently mtu=1500 is the CEBAF default. The addition of the line is
noted in the dip -t -v output stream fragment quoted below (-v, echo on)

>> print I think the mtu is $mtu <- print of $mtu local default
I think the mtu is 296 <- what is printed
>> get $mtu remote 3 <- try to get JLAB end value
About to set variable $mtu to 1500 <- response locally to query
>> print The maybe updated mtu is $mtu <- print updated value to see
The maybe updated mtu is 1500 <- what is printed

The addition of this line allowed me to get through handily. I may try
my hand at Arne's problem in the next day or so. The docs are so much fun
to read, you know.

I didn't have to do _this_ earlier. Maybe it is a consequence of the
masquerade process? I doubt it, and I expect to find it is either another
brain fade issue or some irreproducible coincidence. Sorry to have
bothered all of the rest of you. I hope it was worth at least something
for its entertainment value (as in, "Ho, ho, I sure am glad someone else
does that kind of thing, too").

Michael Tiefenback