Home » Infrastructure » Linux » 10g on Fedora core 2
10g on Fedora core 2 [message #99101] Wed, 21 July 2004 23:45 Go to next message
M. Ashtiani
Messages: 1
Registered: July 2004
Junior Member
running Fedora core 2 under celeron,
DBCA failes with end of communication channel.

when i tried to create database by applying scripts one by one, I traced to :

  * installs Jserver without errors
  * at the begining of intermedia script it executes

   dbms_java.Grant_permission( 'SYSTEM', "FileIoPermission', '<<ALL_FILES>>', 'read' )

and this causes end of communication channel. there is a huge log with core dump about objects %joe%_ not binded

I guess this is a permission problem, any ideas?

( I realize i have sysntax errors, I am away from the server and can not copy the exact messages)

p.s.
I've downloaded the latest kernel and re-compiled it for hugetlbfs=y, wireless support, and block devices=Y
Re: 10g on Fedora core 2 [message #99202 is a reply to message #99101] Wed, 20 October 2004 00:25 Go to previous messageGo to next message
akofoed
Messages: 1
Registered: October 2004
Junior Member
I get the exact same error with 10g on a RedHat 3 ES:

Exception signal: 11 (SIGSEGV), code: 1 (Address not mapped to object), addr: 0x371bc1d0, PC: [[0x10fc19c, joe_well_known
_internal()+102]]
Registers:
%eax: 0x011a4d58 %ebx: 0x01483bb4 %ecx: 0xb7301e8b
...
joe_well_known_internal()+116 (0x10fc1aa) jz 0x10fc239
*** 2004-10-19 16:53:13.541
ksedmp: internal or fatal error
ORA-07445: exception encountered: core dump [[joe_well_known_internal()+102]] [[SIGSEGV]] [[Address not mapped to object]] [[0x
371BC1D0]] [[]] [[]]
Current SQL statement for this session:
call dbms_java.grant_permission ('SYSTEM', 'java.io.FilePermission',
'<<ALL FILES>>', 'read')

None of the checks fails during install !
Strange thing is that we have 2 other machines that has no problems installing Oracle 10g on the exact same OS !!!
So only difference is the hardware. I've even checked to see if there should be differences in the libraries installed - no diffs.

I 2 have the latest 2.4.21-20 kernel
Kind rgds.
Re: 10g on Fedora core 2 [message #99210 is a reply to message #99101] Sun, 31 October 2004 15:55 Go to previous messageGo to next message
Kevin Ross
Messages: 3
Registered: October 2004
Junior Member
I was able to install the database without a problem. But, I am getting a very similar error when trying to run repCA on the server. I continually breaks while attempting to load a jar in ultrasearch. Someone help!

Exception signal: 11 (SIGSEGV), code: 1 (Address not mapped to object), addr: 0xd7779fd8, PC: [[0x10d819c, joe_well_known_internal()+102]]
Registers:
%eax: 0x00e88d58 %ebx: 0x0145fbb4 %ecx: 0xd6d80d8b
%edx: 0xd6d80d8b %edi: 0x0027e491 %esi: 0xd6d80d88
%esp: 0xfeed0ca0 %ebp: 0xfeed0cc0 %eip: 0x010d819c
%efl: 0x00210286
joe_well_known_internal()+91 (0x10d8191) mov 0xfffffffc(%ebp),%edx
joe_well_known_internal()+94 (0x10d8194) mov %edx,%esi
joe_well_known_internal()+96 (0x10d8196) and $-8,%esi
joe_well_known_internal()+99 (0x10d8199) mov %esi,0xffffffec(%ebp)
> joe_well_known_internal()+102 (0x10d819c) mov 0xc(%esi,%edi,4),%esi
joe_well_known_internal()+106 (0x10d81a0) cmp %edx,%esi
joe_well_known_internal()+108 (0x10d81a2) jz 0x10d8294
joe_well_known_internal()+114 (0x10d81a8) test %esi,%esi
joe_well_known_internal()+116 (0x10d81aa) jz 0x10d8239
*** 2004-10-30 00:18:51.563
ksedmp: internal or fatal error
ORA-07445: exception encountered: core dump [[joe_well_known_internal()+102]] [[SIGSEGV]] [[Address not mapped to object]] [[0xD7779FD8]] [[]] [[]]
Current SQL statement for this session:
create or replace java resource named "META-INF/MANIFEST.MF"using blob LOB from CREATE$JAVA$LOB$TABLE where name ='LOADJAVA/LOB/KEY'
Re: 10g on Fedora core 2 [message #99211 is a reply to message #99101] Sun, 31 October 2004 16:42 Go to previous messageGo to next message
Kevin Ross
Messages: 3
Registered: October 2004
Junior Member
I'm using kernel 2.6.6-1.435.2.3 and I'm having the same problem, what version are you using?

-Kevin
Re: 10g on Fedora core 2 [message #99213 is a reply to message #99202] Mon, 01 November 2004 08:09 Go to previous messageGo to next message
Bob Erdman
Messages: 1
Registered: November 2004
Junior Member
I have had the same issue on RH3.0ES, it is the new kernel you are using. Go back to the previous kernel from update1 disks and it will work (2.4.21-15.0.2.ELsmp) I have a tar open and Oracle has replicated it.
Re: 10g on Fedora core 2 [message #99232 is a reply to message #99213] Fri, 19 November 2004 12:54 Go to previous message
Alex Rext
Messages: 1
Registered: November 2004
Junior Member
What's the TAR number?

TIA,
Alex
Previous Topic: Core dump error in Linux
Next Topic: SQL Plus error
Goto Forum:
  


Current Time: Fri Apr 19 20:18:53 CDT 2024