Lots of seg faults

[Scott Burton]

Anyone care go guess what's going on here?

This started happening out of the blue. I've since updated Ruby but
still get "[BUG] Segmentation fault" on most operations involving ruby,
gem, rake, etc.

This is just an example, most every usage of gem or rake will do this.
They fail repeatedly until they work:

------------------------------------------------------

[xserve:/var/www/rails] admin% sudo gem install chronic
Password:
Updating metadata for 80 gems from http://gems.rubyforge.org
................................................................................
complete
Successfully installed rubyforge-0.4.4
Successfully installed hoe-1.4.0
Successfully installed chronic-0.2.3
3 gems installed
Installing ri documentation for rubyforge-0.4.4...
/usr/local/lib/ruby/1.8/rdoc/parsers/parse_rb.rb:489: [BUG] Segmentation
fault
ruby 1.8.6 (2007-09-24) [powerpc-darwin8.11.0]

Abort
[xserve:/var/www/rails] admin% sudo gem install chronic
Successfully installed chronic-0.2.3
1 gem installed
Installing ri documentation for chronic-0.2.3...
/usr/local/lib/ruby/1.8/irb/slex.rb:72: [BUG] Segmentation fault
ruby 1.8.6 (2007-09-24) [powerpc-darwin8.11.0]

Abort
[xserve:/var/www/rails] admin% sudo gem install chronic
Successfully installed chronic-0.2.3
1 gem installed
Installing ri documentation for chronic-0.2.3...
/usr/local/lib/ruby/1.8/irb/slex.rb:76: [BUG] Segmentation fault
ruby 1.8.6 (2007-09-24) [powerpc-darwin8.11.0]

Abort
[xserve:/var/www/rails] admin% sudo gem install chronic
Successfully installed chronic-0.2.3
1 gem installed
Installing ri documentation for chronic-0.2.3...
/usr/local/lib/ruby/1.8/rdoc/parsers/parse_rb.rb:1477: [BUG]
Segmentation fault
ruby 1.8.6 (2007-09-24) [powerpc-darwin8.11.0]

Abort
[xserve:/var/www/rails] admin% sudo gem install chronic
Successfully installed chronic-0.2.3
1 gem installed
Installing ri documentation for chronic-0.2.3...
/usr/local/lib/ruby/1.8/rdoc/parsers/parse_rb.rb:384: [BUG] Segmentation
fault
ruby 1.8.6 (2007-09-24) [powerpc-darwin8.11.0]

Abort
[xserve:/var/www/rails] admin% sudo gem install chronic
Successfully installed chronic-0.2.3
1 gem installed
Installing ri documentation for chronic-0.2.3...
Installing RDoc documentation for chronic-0.2.3...
/usr/local/lib/ruby/1.8/rdoc/parsers/parse_rb.rb:122: [BUG] Segmentation
fault
ruby 1.8.6 (2007-09-24) [powerpc-darwin8.11.0]

Abort
[xserve:/var/www/rails] admin% sudo gem install chronic
/usr/local/lib/ruby/site_ruby/1.8/rubygems/source_index.rb:228: [BUG]
Segmentation fault
ruby 1.8.6 (2007-09-24) [powerpc-darwin8.11.0]

Abort
[xserve:/var/www/rails] admin% sudo gem install chronic
Successfully installed chronic-0.2.3
1 gem installed
Installing ri documentation for chronic-0.2.3...
Installing RDoc documentation for chronic-0.2.3...
[xserve:/var/www/rails] admin%

 

[Thomas Adam]

Hello --

Anyone care go guess what's going on here?

This started happening out of the blue. I've since updated Ruby but
still get "[BUG] Segmentation fault" on most operations involving ruby,
gem, rake, etc.

This is just an example, most every usage of gem or rake will do this.
They fail repeatedly until they work:

Anything else segfaulting? Might be faulty memory -- see memtest86
and friends. If you set:

ulimit -c unlimited

Can you get a backtrace from the resulting corefile created via gdb?

-- Thomas Adam

 

[Scott Burton]

Hello --

Anyone care go guess what's going on here?

This started happening out of the blue. I've since updated Ruby but
still get "[BUG] Segmentation fault" on most operations involving ruby,
gem, rake, etc.

This is just an example, most every usage of gem or rake will do this.
They fail repeatedly until they work:

Anything else segfaulting? Might be faulty memory -- see memtest86
and friends. If you set:

ulimit -c unlimited

Can you get a backtrace from the resulting corefile created via gdb?

-- Thomas Adam

Thanks Thomas,

Here are two backtraces:

-----------
This GDB was configured as "powerpc-apple-darwin".
#0 0x900481ac in ?? ()
(gdb) bt
#0 0x900481ac in ?? ()
#1 0x9012d7b4 in ?? ()
warning: Previous frame identical to this frame (corrupt stack?)
#2 0x0003aaf4 in ?? ()
#3 0x0006c1ac in ?? ()
#4 0x90130f54 in ?? ()
#5 0x0000e3d8 in ?? ()
#6 0x00000000 in ?? ()
Cannot access memory at address 0x42000226
#7 0x00000000 in ?? ()
Cannot access memory at address 0x42000226
Previous frame inner to this frame (corrupt stack?)
-----------


-----------
This GDB was configured as "powerpc-apple-darwin".
#0 0x900481ac in ?? ()
(gdb) bt
#0 0x900481ac in ?? ()
#1 0x9012d7b4 in ?? ()
warning: Previous frame identical to this frame (corrupt stack?)
#2 0x0003aaf4 in ?? ()
#3 0x0006c1ac in ?? ()
#4 0x90130f54 in ?? ()
-----------



"Cannot access memory" looks suspiciously like bad RAM....

Unfortunately I don't have memtest on this box.

 

[Thomas Hurst]

"Cannot access memory" looks suspiciously like bad RAM....

No, that just means the value on the stack isn't a valid memory address
for the process; that, plus the NULLs and the duplicated values in the
backtrace suggest, as gdb says, corrupt stack. The behavior does look
like bad RAM, though.

Unfortunately I don't have memtest on this box.

http://matt.ucc.asn.au/memtest/ any use?

 

[Micah Geisel]

i am having the same problem. was a solution found? did it turn out to
be faulty memory?

 

[Scott Burton]

i am having the same problem. was a solution found? did it turn out to
be faulty memory?

I'll let you know in a few days when I get my RAM

I ran a memtest on it and came back with no faults, but I'm getting
other failures now in Ruby and related processes. Darwin's CrashReporter
core dumps look a whole lot like the one I described above. I can't
prove it but I'm suspicious of the memory, and it's cheap enough to
replace and test it out.

I'll let you know.