topic Re: retry problem in Operating System - HP-UX

retry problem

Etienne Roseau — Wed, 06 Apr 2005 09:23:02 GMT

hi gurus !
i have a k class with a 100base-T card.
everything seem to be OK (ping, lanscan, ioscan) except a msg with nslookup (see attached)
the problem : we have a kind of net guardian with Windows 2000, and, when scanning, it sees "retry" problem with my lancard on my server ????
any suggestion ? is it my problem or windows manager's one ?
thx in advance
E.

Re: retry problem

Pete Randall — Wed, 06 Apr 2005 09:29:52 GMT

Run "lanadmin", then select "lan", then "display" to display the usage statistics. Any errors, discards, collisions? If not, it's the windoze manger's problem.

Pete

Re: retry problem

harry d brown jr — Wed, 06 Apr 2005 09:30:18 GMT

It appears your /etc/nsswitch.conf file is bad.

The info you posted doesn't show any errors with your lan card.

do a lanadmin
lan->display
and post the output of that here

live free or die
harry d brown jr

Re: retry problem

Etienne Roseau — Wed, 06 Apr 2005 09:39:01 GMT

thanks for your fast replies !
please find the lanadmin issue attached
points are coming !
E.

Re: retry problem

Pete Randall — Wed, 06 Apr 2005 09:42:49 GMT

Everything looks good there, except for:

Carrier Sense Errors = 84409

It should be 0 if everything is functioning properly.

Pete

Re: retry problem

Etienne Roseau — Wed, 06 Apr 2005 10:18:52 GMT

OK my mistake :
i change the ppa number in lanadmin and it shows many collisions (see attached)
is there a way for me to find out if the problem is from me (hard or conf ?) or windoze ?
Harry, i checked with my dev server which has no problem, and the same net configuration (except the lancard) and the nsswitch.conf seems to be OK.
THX

Re: retry problem

Pete Randall — Wed, 06 Apr 2005 10:24:37 GMT

This can be because of a duplex mismatch. On the 100base-T cards, the standard recommendation is to hardcode both the card and the port it connects to, to 100FD. You can check the card with "lanadmin -x 0", where 0 is the PPA number of the card. You can change it temporarily using "lanadmin -X 100FD 0". To change it permanently, you need to change the appropriate config file, which depends on the driver for your card (check ioscan):

btlan hpbtlanconf
btlan0 hpeisabtconf
btlan1 hpbasetconf
btlan3 hpbase100conf
btlan4 hpgsc100conf
btlan5 hppci100conf
btlan6 hpsppci100conf

To check the port, you'll need to ask your lan folks.

Pete

Re: retry problem

harry d brown jr — Wed, 06 Apr 2005 10:34:22 GMT

Though not related to your retry's on your lan card, this is troublesome:

nslookup shows :
*** Unknown source : ldap. Not Supported
*** You have defind one or more supported sources.
*** Please modify your configuration file and try again
*** or you will get unreliable information.
Using /etc/hosts on: (hostname)

If you want help with this, post your /etc/nsswitch.conf file:

grep -v "^#" /etc/nsswitch.conf

live free or die
harry d brown jr

Re: retry problem

Ron Kinner — Wed, 06 Apr 2005 10:48:00 GMT

I think Pete has it nailed. You have a 100 Half connection which is sort of unnatural. You are getting a lot of late collisions which should not be the case if you are on a switch unless the other side is set to Full Duplex.

I'd say the other end is manually set to 100 Full and your side is set to AutoNegotiate. Because of the way Autonegotiate is required to work you will always get a mismatch in this situation. (The other end will be seeing a lot of FCS errors and runts.)

You should be able to correct the problem by setting your end to 100 Full. The other end is almost certainly already set to 100 FULL so you don't need to bother the Network guy unless you want to yell at him for screwing up your connection. Once it changes watch the lanadmin stats to make sure you are not getting any

Alignment Errors
or
FCS Errors

which would indicate that it still wasn't correct and now the other end would be set to 100 Half. This would indicate autonegotiation failure so you would have to have the Network guy set the switch to 100 full (or replace the cable).

Ron

Re: retry problem

Etienne Roseau — Thu, 07 Apr 2005 02:55:33 GMT

'morning guys !
for Harry see the doc attached !
for Ron and Pete: I think you got it !!!
actually, my boss told me that the conf was defined to 10 from the windoze lan-manager AND the lanadmin -x (ppa) shows an "autonegociation-ON" !!!
thx everybody
E.

Re: retry problem

Etienne Roseau — Thu, 07 Apr 2005 03:03:22 GMT

OOOOOOPS ! sorry Harry
;-)

Re: retry problem

harry d brown jr — Thu, 07 Apr 2005 06:24:11 GMT

Etienne,

Do you have LDAP installed and configured to resolve host names? If not, then you need to remove the "ldap" from the "hosts" line in /etc/nsswitch.conf.

Do a

swlist LdapUxClient NisLdapServer

to determine if you have the ldap client and/or the nisldap server installed.

live free or die
harry d brown jr

Re: retry problem

Etienne Roseau — Thu, 07 Apr 2005 07:10:14 GMT

interresting......no Ldap utilities found !!
what do you think ??

Re: retry problem

Pete Randall — Thu, 07 Apr 2005 07:13:38 GMT

A properly configured nsswitch entry for hosts should look more like this:

hosts: files[NOTFOUND=continue UNAVAIL=continue] dns [NOTFOUND=continue UNAVAIL=
continue TRYAGAIN=return]

man 4 nsswitch.conf

Pete

Re: retry problem

Etienne Roseau — Thu, 07 Apr 2005 08:17:21 GMT

Harry, Pete,
thx 4 your help with nsswitch.conf.
BUT, my boss told me we don't care about it, because of the net managing by windoze...
our HP is like a client of it !!
> and i don't earn money enough too think about this ;-)

points anyway !!
E.