Luminis 4 Install issue with most recent Solaris 10_Recomended
Has anyone had any issues or attempted to install Luminis 4 Resource Tier with either the most recent ISO of Solaris 10, or any 10_recomended patch cluster since March 20th 2009?
We are in the process of installing a fresh instance of Luminis 4 and have done it successfully a number of times on project machines using Solaris 10 ISO 10/08 with a 10_recomended patch cluster from 3-20-2009 (believe that is the correct date, give or take a day). I attempted to install Luminis 4 in mid May using a Solaris 10_recomended patch cluster from late April and received an error when the installer was configuring the LDAP server. I reverted back to the previous known working Solaris patch revision I had on hand to complete the install. The error is as follows:
Error: [E315] The 'perl' external program failed (exit code = 49): Welcome to the Directory Server preparation tool for Sun Java(tm) System communication services. (Version 6.3 Revision 2.03) This tool prepares your directory server for use by the communications services which include Messaging, Calendar and their components. The logfile is /var/tmp/dssetup_20090522095300.log. The entered password for DN cn=Directory Manager is incorrect. Please correct the problem and re-run this script. Remedy: [F000] Contact Sungard Higher Education Support
This error also was present using the most recent 10_recomended patch cluster from 6-18-09.
The log files don't have any insight into the cause of the problem and the only error messages state that the password entered is incorrect. The same install.conf file was used from previous successful Luminis 4 Installs, except for hostname changes.
Also, if I install the Luminis 4 resource tier first then patch, it seems to break the directory server. I did not check the logs for that attempt due to project deadlines to have a working system in place and just reverted to what worked before.
Our SCT support ticket was closed when I stated I was installing using a previous patch cluster and no mention was given that they would investigate.
Anyone else ran into a similar issue?

Most Recent Solaris_10 install
I did one using the March 2009 release, sysadmin says the patches, that iPatch said all were up-to-date (re-loaded in May sometime)
The only issue I had was with the install binary being read correctly when mounted in a zone as zfs vs ufs - the error wasnt plain and was during the initial JES install. Once I remounted all went as usual. I script as much as possible of the pre-install "Known Issues" so very little is left to chance.
Sounds like the 10_rec patches are breaking things, I'd go with U7 : my uname >
Generic_139555-08 sun4v sparc SUNW,Sun-Blade-T6300
good luck
Solaris patch 119213-19 is the culprit
We ran into this problem and it is caused by Solaris patch 119213-19. This patch can removed or You can set the following environment variable in /etc/profile.
NSS_STRICT_NOFORK=DISABLED
It is documented in the Luminis Platform IV known issues doc but it is hard to find.
Mike
Thanks, that was it. The
Thanks, that was it. The Known issues document that I have doesn't have anything about that issue there. Searched Sungard's site and finially found the updated one.
Thanks again,
Chris
Interesting...
I had come to the conclusion it was something breaking SSHA passwords. The problem I had was all authentication to LDAP breaking. In troubleshooting I edited the dse.ldif file and changed the Directory Manager password from a {SSHA} to a {CRYPT} hash to see if it would enable that user's password. It did indeed work, resetting the value to another {SSHA} generated value did not. I suspected and confirmed this patch was the culprit after reviewing the installed patches and noting this one was the only one remotely having anything to do with Sun Directory Services.
Here's a good thread on Sun about the issue:
http://forums.sun.com/thread.jspa?threadID=5388655
Also from here the patch notes themselves:
https://sunsolve.sun.com/search/document.do?assetkey=119213
This seems suspicious:
Perhaps the changes to support higher levels of SHA encoding broke salted sha?
Nope, don't think so
It sounds like you may not have read the bottom of the patch description:
It is by design that LDAP doesn't work right after this patch. The only thing that is a little unfortunate is that there is no patch to the directory server that fixes the behavior, only a workaround (unless anyone else reads this differently than I).
FYI, I got around this by setting the parameter in my .cprc file, not /etc/profile -- I really didn't want everything bypassing that rule, if possible, only necessary products.