Endian wierdness

mfrench42 · n00b Joined: 05 Jan 2004 Posts: 3 Location: South Africa

Hi all,

I have recently installed Gentoo on my Ultra5. It was previously running Debian without any problems. But I have encountered a wierd issues.

I was trying to get LDAP up and running. The ldap tools compiled locally could access the database without a hitch. But remote tools compiled on Intel platforms could not. The logs reported an error related to "number of out of range". This smells to me like a classic endianess problem.

The PAM/LDAP module on the local system could not connect to the LDAP database either.

My CFLAGS is not very unusual. I only have the UltraSparc optimisation set:
CFLAGS="-mcpu=v9 -mtune=v9 -O2 -pipe"

This is not serious, but I wonder if anyone else has encountered this problem?

- Matthew

zappa_the_great · Posted: Tue Jan 06, 2004 12:00 am Post subject:

Hi,

works fine here (LDAP server on Ultra 1, remote clients running Linux/x86, Irix and pam_ldap/nss_ldap). Which version of OpenLDAP did you emerge? I'm using 2.1.23 (hda to accept "~sparc"). Did you try to run your searches with full debugging flags on both server and client? You could also switch on slapd logging via syslog, maybe there's something in the log files.

JM
_________________
"I haven't lost my mind -- it's backed up on tape somewhere."

mfrench42 · n00b Joined: 05 Jan 2004 Posts: 3 Location: South Africa

Drat. The last thing I want to hear is that someone else has no problem...

It is version 2.1.23, and I have run it with full debugging flags. Don't have it right now, and the LDAP server appears to have lost its schema. But AFAIR the error message on the client side was a sudden disconnect and on the server side it was a protocol error - number out of range.

I will have to emerge again, rebuild the database, and then see if the problem persists.

- Matthew

zappa_the_great · Posted: Tue Jan 06, 2004 4:12 pm Post subject:

Hi,

was it exactly "number out of range" or rather "numerical result out of range"? The OpenLDAP FAQ has these pointers:

Protocol Error:

http://www.openldap.org/faq/data/cache/693.html

Result out of range:

http://www.openldap.org/faq/data/cache/671.html

Did you try to force LDAPv2 on the client side? I had all kind of weird errors too because of my clients trying to negotiate v3. Also, TLS/SSL sometimes seems to fail, so you might have to disable it on both sides or set to start_tls.

JM
_________________
"I haven't lost my mind -- it's backed up on tape somewhere."

mfrench42 · n00b Joined: 05 Jan 2004 Posts: 3 Location: South Africa

Well, I recompiled and the same problem cropped up. But then I tried using my new gentoo x86 laptop and it could query the server without hassles.

The laptop was using RedHat 8.0 and the other clients are running RedHat 9.0. RPM reports OpenLdap version 2.0. 27 on these machines. But these seems to be the problem. :oops:

FYI: the FAQ question I needed to find (but did not) was:

http://www.openldap.org/faq/data/cache/822.html

*sigh* I was reluctant to bump the RedHat workstations to Debian because of bad experiences with unstable. But so far Gentoo has allowed me to ride the upgrade whirlwind without breaking anything. I guess it might be time for the annual workstation reformat...

- Matthew