On 10/5/2004 3:35 PM, Brad Hilton wrote:

> Jan Wieck wrote:
>> On 10/4/2004 8:30 PM, Brad Hilton wrote:
>> 
>>> Hello,
>>>
>>> I've spent the last few days learning slony and setting it up on some
>>> test databases with good results.  However, when I set up replication on
>>> my live database, the slon processes startup and appear to be preparing
>>> to copy things over when I get:
>>>
>>> "The postmaster has commanded this server process to roll back the
>>> current transaction and exit, because another server process exited
>>> abnormally and possibly corrupted shared memory."
>> 
>> 
>> Could you provide a stack trace of the crashed Postgres process, please?
> 
> OK, here it is:

That is inside the Slony log trigger, at the place where it compares the 
two values of an attribute from OLD and NEW, so the operation in 
question was an UPDATE. Now both values are NOT NULL, and the attribute 
is not marked dropped in the tuple descriptor. What the trigger tries to 
  do is calling the equality comparision function for the corresponding 
data type and the function call info was obtained via lookup_type_cache. 
And I have no clue yet why THAT would have failed.

In that debug session, could you please do:

     frame 2
     print *type_cache

Another question is "what got dropped to cause this"? I can only assume 
that some some data type definition changes or column type changes have 
been invoked in this. But this is pure guessing.


Thank you for the time debugging this
Jan

> 
> #0  0x00000000 in ?? ()
> #1  0x081eabfd in FunctionCall2 (flinfo=0x8313870, arg1=3221211232,
>      arg2=3221211232) at fmgr.c:993
> #2  0x4001a8cc in _Slony_I_logTrigger (fcinfo=0x20) at slony1_funcs.c:600
> #3  0x081ea27f in fmgr_security_definer (fcinfo=0xbfffca60) at fmgr.c:686
> #4  0x080f62de in ExecCallTriggerFunc (trigdata=0xbfffcbb0, finfo=0x83c59b0,
>      per_tuple_context=0xbfffc860) at trigger.c:1150
> #5  0x080f7234 in DeferredTriggerExecute (event=0x8382830, 
> itemno=-1073756064,
>      rel=0x24, trigdesc=0x83057e0, finfo=0x8313870,
>      per_tuple_context=0xbfffc860) at trigger.c:1859
> #6  0x080f763d in deferredTriggerInvokeEvents (immediate_only=1 '\001')
>      at trigger.c:2000
> #7  0x080f77b0 in DeferredTriggerEndQuery () at trigger.c:2135
> #8  0x0817cbd8 in finish_xact_command () at postgres.c:1749
> #9  0x0817b8f3 in exec_simple_query (
>      query_string=0x83039c0 "insert into comments (imported, 
> convert_newlines, thread_sortorder, depth, parent_id, article_id, 
> user_id, id, formatted_body, formatted_title, body, remote_ip, 
> author_url, author_email, author_name, "...)
>      at postgres.c:905
> #10 0x0817dbb0 in PostgresMain (argc=4, argv=0x82ba358,
>      username=0x82ba328 "blogware") at postgres.c:2868
> #11 0x08157a7a in BackendFork (port=0x82cbcf8) at postmaster.c:2564
> #12 0x08157483 in BackendStartup (port=0x82cbcf8) at postmaster.c:2207
> #13 0x08155a38 in ServerLoop () at postmaster.c:1119
> #14 0x0815532a in PostmasterMain (argc=3, argv=0x82b9df8) at 
> postmaster.c:897
> #15 0x08125367 in main (argc=3, argv=0xbfffdd34) at main.c:214
> 
> ----------------------
> Here's flinfo from frame 1 in case it helps:
> 
> p *flinfo
> $2 = {fn_addr = 0, fn_oid = 0, fn_nargs = 0, fn_strict = 0 '\0',
>    fn_retset = 0 '\0', fn_extra = 0x0, fn_mcxt = 0x0, fn_expr = 0x0}
> 
> Let me know what if there's anything else I can provide.
> 
> Thanks,
> Brad
> _______________________________________________
> Slony1-general mailing list
> Slony1-general at gborg.postgresql.org
> http://gborg.postgresql.org/mailman/listinfo/slony1-general


-- 
#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me.                                  #
#================================================== JanWieck at Yahoo.com #