On Thu, Mar 11, 2010 at 2:59 PM, mark <prg...@gmail.com> wrote:
> On Thu, Mar 11, 2010 at 9:25 AM, Kenny Lussier <kluss...@gmail.com> wrote:
>>
>> Hi all,
>>
>> I have the unfortunate need to reproduce a server that was built 6
>> years ago, and make them identical. The server is RHEL3 i386. I have
>> managed to get the boxes to an identical state at the OS and package
>> level, and everything seems to work. However, there is one thing that
>> has me puzzled. On the original box, when a child process is forked,
>> it is hidden from `ps`. In one case, if I do a `ps auxww | grep
>> splunk`, I get:
>>
>> I have read up on this, and I understand group leaders, and group
>> member non-leaders. The weirdness comes in on the new system. Exact
>> same kernel, package-for-package identical to the first. The
>> difference is that there are no .pid files in /proc, and ps shows
>> every child:
>>
>
> The hidden files in /proc are threads that have gotten forked, see man 2
> clone for more on that.  You can see these in ps using the -m option:  ps
> -efm to get the listing; and ps-efm|sort -n -k2 to see them sorted by PID,
> which may be more helpful to see what spawned what.
>
> Look at the libc and glibc rpms on both systems to see if they are
> identical. To be certain, check the --info data for each and see if the
> build date, build host, and signature are all identical; e.g.:
>
>  rpm -q glibc-headers-2.3.2-95.33 --info
> Name        : glibc-headers                Relocations: (not relocatable)
> Version     : 2.3.2                             Vendor: Red Hat, Inc.
> Release     : 95.33                         Build Date: Wed 23 Feb 2005
> 08:01:07 AM EST
> Install Date: Tue 28 Jun 2005 01:05:53 PM EDT      Build Host:
> porky.build.redhat.com
> Group       : Development/Libraries         Source RPM:
> glibc-2.3.2-95.33.src.rpm
> Size        : 1915921                          License: LGPL
> Signature   : DSA/SHA1, Tue 29 Mar 2005 01:28:25 PM EST, Key ID
> 219180cddb42a60e
> Packager    : Red Hat, Inc. <http://bugzilla.redhat.com/bugzilla>
> Summary     : Header files for development using standard C libraries.
> Description :
> The glibc-headers package contains the header files necessary for
> developing programs which use the standard C libraries.  If you are
> developing programs which will use the standard C libraries, your system
> needs to have these standard header files available in order to create the
> executables.
>
> Install glibc-headers if you are going to develop programs which will
> use the standard C libraries.
>
> There may be something under the /proc/sys subdirectories that may have a
> direct or indirect influence on this too.
>
> I'd also diff the /etc/sysctl.conf files, 'cause I'm paranoid like that.

I've done diffs of /etc on them, and the only differences are things
that have to be different (ip address, etc.). sysctl, /etc/security/*
and /boot/config-* were the first things I checked.  The mystery has
gotten stranger, too. There is apparently a 2nd server that is now
displaying the same behavior. The system works, and almost everything
seems to be functioning fine. The only problem that I have seen so far
is that Tripwire scans are failing due to java memory issues. I'm not
sure if it's related or not yet.

I'll be doing some experimenting tomorrow (and probably all weekend)
to find out what went wrong.

Thanks,
Kenny

_______________________________________________
gnhlug-discuss mailing list
gnhlug-discuss@mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/

Reply via email to