Hi,
Right now pg_waldump just prints whether the message is transactional
or not and its size. That doesn't help much to understand the message
itself. If it prints the contents of a logical WAL message, it helps
debugging logical replication related problems. Prefix is a
null-terminated ASCII string, so no problem printing that. Even the
contents can be printed as a series of hex bytes. Here's a patch to do
that.
I tested this manually as below
postgres=# select pg_logical_emit_message(false, 'some_prefix', 'some
message'::text);
pg_logical_emit_message
-------------------------
0/1570658
(1 row)
$> pg_waldump --start 0/1570600 -p data/
first record is after 0/1570600, at 0/1570608, skipping over 8 bytes
rmgr: LogicalMessage len (rec/tot): 74/ 74, tx: 0,
lsn: 0/01570608, prev 0/015705D0, desc: MESSAGE nontransactional
message size 12 bytes, prefix some_prefix; mesage: 73 6F 6D 65 20 6D
65 73 73 61 67 65
rmgr: Standby len (rec/tot): 50/ 50, tx: 0, lsn:
0/01570658, prev 0/01570608, desc: RUNNING_XACTS nextXid 504
latestCompletedXid 503 oldestRunningXid 504
pg_waldump: fatal: error in WAL record at 0/1570658: invalid record
length at 0/1570690: wanted 24, got 0
I didn't find any tests for pg_waldump to test its output, so haven't
added one in the patch.
--
Best Wishes,
Ashutosh Bapat
