Re: BUG #15570: Vacuum analyze ERROR: MultiXactId XXXX has not been created yet -- apparent wraparound

From: Andrew Gierth <andrew(at)tao11(dot)riddles(dot)org(dot)uk>
To: Phil Hildebrand <phil(dot)hildebrand(at)moz(dot)com>
Cc: Alvaro Herrera <alvherre(at)2ndquadrant(dot)com>, pgsql-bugs(at)lists(dot)postgresql(dot)org
Subject: Re: BUG #15570: Vacuum analyze ERROR: MultiXactId XXXX has not been created yet -- apparent wraparound
Date: 2019-01-01 06:56:58
Message-ID: 87muoksx17.fsf@news-spur.riddles.org.uk
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

>>>>> "Phil" == Phil Hildebrand <phil(dot)hildebrand(at)moz(dot)com> writes:

Phil> Sure - beacon=# select lp, lp_flags, t_xmin, t_xmax, t_field3,
Phil> t_ctid, beacon-# to_hex(t_infomask) as infomask,
Phil> to_hex(t_infomask2) as infomask2 beacon-# from
Phil> heap_page_items(get_raw_page('reviews_2018', 382604)) h;

This data page has obviously been partially overwritten by unrelated
data. (Notice that many of those numbers in the output correspond to
printable ASCII strings.)

A hexdump of the raw page will probably make it easy to see which part
is corrupt and what the offending data is. Something like this is
probably the easiest way to get it (adding any other required psql
options of course):

psql -d beacon -AXqt -c "select get_raw_page('reviews_2018', 382604)" |
perl -ne '/\\x([[:xdigit:]]+)/ and print pack("H*",$1)' |
hexdump -C

Assuming that doesn't reveal any sensitive data, can you send the output
of that?

--
Andrew (irc:RhodiumToad)

In response to

Responses

Browse pgsql-bugs by date

  From Date Subject
Next Message Phil Hildebrand 2019-01-01 07:25:21 Re: BUG #15570: Vacuum analyze ERROR: MultiXactId XXXX has not been created yet -- apparent wraparound
Previous Message Phil Hildebrand 2019-01-01 01:09:25 Re: BUG #15570: Vacuum analyze ERROR: MultiXactId XXXX has not been created yet -- apparent wraparound