Re: vacuum analyze fails with latest cvs version - Mailing list pgsql-bugs

From Jukka Honkela
Subject Re: vacuum analyze fails with latest cvs version
Date
Msg-id Pine.GSO.4.30.0101012252540.13899-100000@stekt1
Whole thread Raw
In response to Re: vacuum analyze fails with latest cvs version  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-bugs
On Mon, 1 Jan 2001, Tom Lane wrote:

>pgsql-bugs@postgresql.org writes:
>> Problem 1: "vacuum messages;" works but "vacuum analyze messages;"
>> kills the backend.
>
>We have absolutely no hope of responding to a bug report with only
>that amount of information.  How about a backtrace from the core
>file?  How about telling us the schema for the table that's causing
>the problem?


(gdb) bt
#0  0x40269e80 in strcoll () at strcoll.c:228
#1  0x812301a in varstr_cmp ()
#2  0x8123067 in text_cmp ()
#3  0x812309a in text_lt ()
#4  0x8135036 in FunctionCall2 ()
#5  0x80acffd in attr_stats ()
#6  0x80acd19 in analyze_rel ()
#7  0x80a91fe in vac_vacuum ()
#8  0x80a916d in vacuum ()
#9  0x80fc06f in ProcessUtility ()
#10 0x80fa119 in pg_exec_query_string ()
#11 0x80fb02a in PostgresMain ()
#12 0x80e64e8 in DoBackend ()
#13 0x80e60b7 in BackendStartup ()
#14 0x80e52e5 in ServerLoop ()
#15 0x80e4ce6 in PostmasterMain ()
#16 0x80c6818 in main ()



Table "messages"
Attribute |   Type    | Modifier
mid       | integer   | not null default nextval('msg_id'::text)
lid       | integer   |
sender    | text      |
date      | timestamp |
subject   | text      |
body      | text      |
Index: messages_index



Index "messages_index"
Attribute |  Type
-----------+---------
lid       | integer
mid       | integer
btree


select count(*) from messages;
 count
-------
 18667


select max(length(body)) from messages;
  max
-------
 92148
(1 row)

This is from the stderr of the backend: (it dies on "messages" even
it doesn't show up here)

DEBUG:  Index pg_toast_18752_idx: Pages 34; Tuples 7681. CPU 0.05s/0.03u
sec.
DEBUG:  Analyzing...
Server process (pid 9850) exited with status 139 at Mon Jan  1 01:43:08
2001
Terminating any active server processes...
Server processes were terminated at Mon Jan  1 01:43:08 2001

I didn't compile the thingy with --enable-debug, I can try with that
too, if required?

...
Jukka Honkela

pgsql-bugs by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: PostgreSQL 7.0.3: Memory leak in ESQL library
Next
From: "Denis V. Osadchy"
Date:
Subject: Output of date_part('quarter', date)