I have a recover situation related to:
Oct 13 23:04:58 66-162-145-116 postgres[16955]: [1-1] LOG: database
system was shut down at 2007-10-13 23:04:54 PDT
Oct 13 23:04:58 66-162-145-116 postgres[16955]: [2-1] LOG: checkpoint
record is at F0/E200001C
Oct 13 23:04:58 66-162-145-116 postgres[16955]: [3-1] LOG: redo record
is at F0/E200001C; undo record is at F0/E200001C; shutdown TRUE
Oct 13 23:04:58 66-162-145-116 postgres[16955]: [4-1] LOG: next
transaction ID: 172668192; next OID: 88470513
Oct 13 23:04:58 66-162-145-116 postgres[16955]: [5-1] LOG: next
MultiXactId: 32334; next MultiXactOffset: 69955
Oct 13 23:04:58 66-162-145-116 postgres[16955]: [6-1] PANIC: could not
access status of transaction 172668192
Oct 13 23:04:58 66-162-145-116 postgres[16955]: [6-2] DETAIL: could not
open file "pg_clog/00A4": No such file or directory
Oct 13 23:04:58 66-162-145-116 postgres[16953]: [1-1] LOG: startup
process (PID 16955) was terminated by signal 6
Oct 13 23:04:58 66-162-145-116 postgres[16953]: [2-1] LOG: aborting
startup due to startup process failure
~
Based on what I've read on the mail archives, the recommended fix is to
create file '00A4' and fill it with 256k zeros. Is there a quick and
easy linux-way of creating such a beast?
-jason
