traceroute hangs from one specific server

[Ron Kinner]

Tom,

Just wanted you to know we have an 11.0 machine in our with the same problem. Just hangs instead of giving * like it should. I'm positive it's a software issue since it hangs when it traces to the default router on the same hub. (Cisco's always fail the second trace since they don't like to respond to too many ICMP messages in too short a time.)

However, this 11.0 machine does have patch 23456 applied so that's not it.

uname -a tells me it's a 785 running 11.0. The production machine which works is an 800 running 11.0. It also show that it has two ICMP patches:
20444 and 20446 which the lab machine does not have. Don't know if it's significant. The hpux boxes are not mine - I just have a login on them so I can't experiment.

Ron

[Tom Geudens]

Hi Ron,
Well, at least I'm glad I'm not alone ;-). Both PHSS_20444 and PHSS_20446 are on all my sytems (which are both N4000 running HP-UX 11.0) ... so those are not the problem either. When I say PHNE_23456, I actually mean PHNE_26771 (the latest and recommended patch) which I had to install for Veritas Netbackup 4.5. It's just that anything above PHNE_23456 will kill Control-M. However, you could be right and the ARP-"line of thought" may also be wrong. Any change you could install PHNE_26771 on that system ?

Regards,
Tom

[Ron Kinner]

I'll ask them but they are a separate company so not much hope.

Do you have PHNE_21835, PHNE_20125 and PHNE_21888? These are for INETD, NDD and LIBNM. The working system has these but the broken one doesn't.

There are actually a bunch of patches which are missing on the broken one but most of them do not appear to be important.

Ron

[Tom Geudens]

Hi Ron,
That's a yes on PHNE_20125 and a no on the other patches. However, this is the same on both systems (the working and the faulty one).

Regards,
Tom

P.S. I'm taking the next week off (starting from today) so although I'd appreciate lots of answers/thoughts I can't answer or assign points for a week.

[Steven E. Protter]

Your problem may be a router.

I have had issues with older Cisco routers with out of date IOS versions not routing properly. Sure as the sun rises, this will cause a problem looking very much the same as your original symptom.

[nick carter_1]

tom geudens!!
is that you?? i have spent years searching for you , if it is. Linz 96??
if it isnt then sorry, if it is my mensafest amigo. email me properly on njc851@hotmail.com "soon" oder big trouble!!

[Ron Kinner]

Tom,

When you get back how about running the following command on each machine with a > junka and > junkb so we will have the info in a file then get both files on the same box then run
diff junka junkb > junkc
and post junkc (or email it to me at rkinner@att.net) and I'll compare my 2 lists and perhaps we can nail down which patches are missing and might be causing the problem.

swlist -l patch|cut -c 8-13|grep -v [a-z,A-Z]|sort -u

This creates a sorted list of every patch (numbers only) on your machine.

Ron

[Tom Geudens]

Ron,
# I executed the following on the non-working server sv00127 and the working server sv00128
swlist -l patch PH* | grep '^# PH' | cut -c8-12 | sort -u

# Then did the diff ...
diff swlist_sv00127.out swlist_sv00128.out > swlist_diff.txt

# The result is in attachment ...

Nick,
Although I've been in the UK on several occasions ... :-( I've never been to Mensafest in Linz (my loss, I know). However, if you're looking for a Belgian "Tom Geudens" I might be able to help you. Is that the case ?

Regards,
Tom

[Tom Geudens]

Ron,
# I executed the following on the non-working server sv00127 and the working server sv00128
swlist -l patch PH* | grep '^# PH' | cut -c8-12 | sort -u

# Then did the diff ...
diff swlist_sv00127.out swlist_sv00128.out > swlist_diff.txt

# The result is in attachment ...

Nick,
Although I've been in the UK on several occasions ... :-( I've never been to Mensafest in Linz (my loss, I know). However, if you're looking for a Belgian "Tom Geudens" I might be able to help you. Is that the case ?

Regards,
Tom

[Ron Kinner]

Amazing number of differences between two systems that were thought to be at about the same patch level. I checked your left arrows against my good box and bad box. Most of them are not on the working system. The only one was 23274 which I know you just installed on the bad box. (You really ought to put it on the good one too.)

I compared the right arrows against both my good box and my bad box (which is missing the 23274) and again you are much advanced over my stuff. I only have 20141, 20145, 22677, 23117, 23409, 24189 on my good box. The bad box has the same except is missing 24189.

Odd. PHCO_24189 is a passwd patch.

http://www2.itrc.hp.com/service/patch/patchDetail.do?patchid=PHCO_24189&context=hpux:800:11:00

Don't see how it could have any effect on traceroute but I will see if I can get my guys to install it and see if it does anything. Don't hold your breath tho. They are very short-handed right now.

Ron