This content has been marked as final. Show 37 replies
Sorry for the late response...1 person found this helpful
How can I verify that I have OS authentication enabled?OS authentication is enabled by default on Linux installs. The parameters you've posted and the absence of an sqlnet.ora indicate that you don't have anything on your install that would disable it.
Is there any problem using the localhost loopback ip address?I don't think so. Either way, IP configuration doesn't matter if you use the local login based on IPC (inter process communication).
All error msgs indicate I do not have sufficient privileges or the userid/passwd are invalid.The privileges require you "just" to be in dba group. If this doesn't work, eventhough your user is a member of that group, something is really odd on your OS install...
How do you login your "oracle" user? And have you tried to create a "fresh" account that is member of dba group?
I noticed that the, lsnrctl status command, indicates that security is ON: Local OS Authentication. So, does that mean OS authentication is not working?No, it means OS authenticaion IS enabled...
Which is probably why the password you entered doesn't work. The root cause for this is the insufficient privileges error for the attempt to use OS authentication in the installer script.
The lines above occurred for each SQL attempt to set a passwd.
Configuring database...ERROR: ORA-01031: insufficient privileges SP2-0306: Invalid option.
There are only a few options left. The most likely ones include some unfullfilled requirements or some unwanted parameters set in your install shell or system. Could you check what environment variables are set for your "root" user when you run the XE installer?
Could it be you have more than one group named dba on your system? To check, you could execute
cut -f1 /etc/group -d ":" | sort | uniq -d
Edited by: Udo on 30.01.2013 16:00
corrected chain for duplicate check
oracle 7562 1 0 06:30 ? 00:00:00 xe_w000_XE
root 7671 31022 0 06:35 pts/4 00:00:00 grep oracle
root 24890 24878 0 Jan24 pts/2 00:00:00 su - oracle
oracle 24891 24890 0 Jan24 pts/2 00:00:00 -bash
oracle 26205 1 0 Jan25 ? 00:00:02 xe_pmon_XE
oracle 26209 1 0 Jan25 ? 00:00:01 xe_psp0_XE
oracle 26214 1 0 Jan25 ? 00:00:00 xe_vktm_XE
oracle 26220 1 0 Jan25 ? 00:00:00 xe_gen0_XE
oracle 26224 1 0 Jan25 ? 00:00:00 xe_diag_XE
oracle 26228 1 0 Jan25 ? 00:00:00 xe_dbrm_XE
oracle 26232 1 0 Jan25 ? 00:00:05 xe_dia0_XE
oracle 26236 1 0 Jan25 ? 00:00:00 xe_mman_XE
oracle 26240 1 0 Jan25 ? 00:00:01 xe_dbw0_XE
oracle 26244 1 0 Jan25 ? 00:00:02 xe_lgwr_XE
oracle 26248 1 0 Jan25 ? 00:00:06 xe_ckpt_XE
oracle 26252 1 0 Jan25 ? 00:00:05 xe_smon_XE
oracle 26256 1 0 Jan25 ? 00:00:00 xe_reco_XE
oracle 26260 1 0 Jan25 ? 00:00:10 xe_mmon_XE
oracle 26264 1 0 Jan25 ? 00:00:12 xe_mmnl_XE
oracle 26268 1 0 Jan25 ? 00:00:00 xe_d000_XE
oracle 26272 1 0 Jan25 ? 00:00:00 xe_s000_XE
oracle 26276 1 0 Jan25 ? 00:00:00 xe_s001_XE
oracle 26280 1 0 Jan25 ? 00:00:00 xe_s002_XE
oracle 26284 1 0 Jan25 ? 00:00:00 xe_s003_XE
oracle 26401 1 0 Jan25 ? 00:00:00 xe_vkrm_XE
oracle 26405 1 0 Jan25 ? 00:00:00 xe_qmnc_XE
oracle 26433 1 0 Jan25 ? 00:00:05 xe_cjq0_XE
oracle 26481 1 0 Jan25 ? 00:00:00 xe_q000_XE
oracle 26485 1 0 Jan25 ? 00:00:00 xe_q001_XE
oracle 26581 1 0 Jan25 ? 00:00:00 xe_smco_XE
root 26760 16920 0 Jan25 pts/3 00:00:00 su - oracle
oracle 26761 26760 0 Jan25 pts/3 00:00:00 -bash
oracle 29858 1 0 Jan25 ? 00:00:01 /u01/app/oracle/product/11.2.0/xe/bin/tnslsnr LISTENER -inherit
Udo, the hint regarding multiple dba groups, in /etc/group, was the issue.
I have this huge /etc/group file where one line entry might be many, many lines long. It did have both the original dba entry, which was difficult to see as the line was so long you weren't seeing the beginning of the line and the dba line entry I added.
Although I tried removing the first dba entry from the /etc/group file and restarted the Linux image, that didn't seem to work. I'm not sure if the orapwd was causing any problems or not, but removing the oracle-xe RPM and reinstalling it, I was able to login to Oracle via sqlplus. Prior to executing the "oracle-xe configure" I removed the standard error and standard out redirection from the end of the script where it creates the sys and system passwds. This time, there were no failure codes to standard out. I was pretty sure at that point I would be able to successfully login to sqlplus.
This was really, really tough to find, but your hint, solved the problem. Many, many thanks Udo!!!
-bash-3.2$ sqlplus / as sysdba
SQL*Plus: Release 184.108.40.206.0 Production on Mon Jan 28 08:54:36 2013
Copyright (c) 1982, 2011, Oracle. All rights reserved.
Oracle Database 11g Express Edition Release 220.127.116.11.0 - 64bit Production
Thanks so much!
I'm glad to hear we've finally found the cause for our headaches...
Especially because this really has been a long struggle, it would be great if other users searching for similar issues could find the essential pieces in this thread easily. To support that, could you please mark any post you think is worth reading as "helpful" (e. g. because it contains a useful hint for troubleshooting) and finally the solution for your issue as "correct"?
You might want to change this command, in your example: cut -f1 /etc/group -d ":" | uniq -d
uniq -d will only catch like values that are adjacent, but not if they are separated by any other value(s).
It would catch:
How about: cut -f1 /etc/group -d ":" | grep dba
I didn't realize that you can only do helpful or correct, but not both. This should have been marked as CORRECT. Is there any way to change this now?
uniq -d will only catch like values that are adjacent, but not if they are separated by any other value(s).That's true, thanks for pointing this out! I took the wrong line from my shell history. The final command I tested with looked like
and of course, you could grep for dba, but I think you should check all group entries for duplicates.
cut -f1 /etc/group -d ":" | sort | uniq -d
P. S.: Thanks for marking the posts! It should be possible to remove the helpful marker, but I'm not sure this (still) works since last update of the forum software. Of course, you could use this post with the ajusted chain for the duplicate check.