Bad records in table

Friends,
    I have a table with bad data in it.  Whenever the backend hits the tupples
with bad data I get:

Jul  2 07:44:18 db postgres[2652]: [22] DEBUG:  server process (pid 2869)
was terminated by signal 11
Jul  2 07:44:18 db postgres[2652]: [23] DEBUG:  terminating any other active
server processes
Jul  2 07:44:18 db postgres[2652]: [24] DEBUG:  all server processes
terminated; reinitializing shared memory and semaphores
Jul  2 07:44:18 db postgres[2871]: [25] FATAL 1:  The database system is
starting up
Jul  2 07:44:18 db postgres[2870]: [25] DEBUG:  database system was
interrupted at 2002-07-02 07:42:30 PDT
Jul  2 07:44:18 db postgres[2870]: [26] DEBUG:  checkpoint record is at
1/94E54768
Jul  2 07:44:18 db postgres[2870]: [27] DEBUG:  redo record is at
1/94E54768; undo record is at 0/0; shutdown TRUE
Jul  2 07:44:18 db postgres[2870]: [28] DEBUG:  next transaction id:
55508992; next oid: 13249044
Jul  2 07:44:18 db postgres[2870]: [29] DEBUG:  database system was not
properly shut down; automatic recovery in progress
Jul  2 07:44:18 db postgres[2870]: [30] DEBUG:  redo starts at 1/94E547A8
Jul  2 07:44:18 db postgres[2870]: [31] DEBUG:  ReadRecord: record with zero
length at 1/94E547CC
Jul  2 07:44:18 db postgres[2870]: [32] DEBUG:  redo done at 1/94E547A8
Jul  2 07:44:20 db postgres[2870]: [33] DEBUG:  database system is ready

    I can identify the bad records with a simple perl script that reads all the
records sequentially and lets me know which ones crash the back end.  Now
that I know which ones they are, how can I get rid of them?

    Once I get this all fixed, what kinds of tests should I do on the machine
to see if it's got some kind of hardware problem?  It's RedHat Linux
2.4.9-31smp.

Thanks for your help!
Peter Darley