<div style="font-family: Verdana; font-size: 12pt; color: #000000"><br /><br /><hr id="zwchr" /><b>From: </b>"Bruce
Momjian"<bruce@momjian.us><br /><b>To: </b>"Lou Picciano" <loupicciano@comcast.net><br /><b>Cc:
</b>pgsql-hackers@postgresql.org<br/><b>Sent: </b>Wednesday, August 31, 2011 10:38:01 PM<br /><b>Subject: </b>Re:
[HACKERS]toast error after pg_upgrade 9.0.4 -> 9.1rc1<br /><br />Lou Picciano wrote:<br />> After running an
essentiallyuneventful* pg_upgrade from 9.0.4 -><br />> 9.1rc1, we are seeing some toast errors logged on the new
cluster:<br/>> <br />> All are of this pattern: ERROR: missing chunk number 0 for toast value<br />> 130087 in
pg_toast_34735<br/>> <br />> Have seen the same pattern for a few of the databases in the 9.1rc1<br />>
cluster,and all as a result of a select on a usr table (the offending<br />> SELECT happens to be the first one any
ofthese DBs sees, as it's the<br />> first step in a user authentication process). SELECT count(*) does not<br
/>>produce an error.<br />> <br />> *almost uneventful: We also saw messages that the destination cluster<br
/>>did not have one of our schema - (of course it didn't!) - I didn't<br />> realize pg_upgrade doesn't 'do'
schema?<br/><br />I have fixed two errors in pg_upgrade since 9.1rc1 was released. Can<br />you use git 9.1 head or
waitfor 9.1rc2 or 9.1 final? I can email you<br />patches if you prefer.<br /><br />--<br /> Bruce Momjian
<bruce@momjian.us> http://momjian.us<br /> EnterpriseDB
http://enterprisedb.com<br/><br /> + It's impossible for everything to be true. +<br /><br />Bruce, many thanks. I've
donethe recent git pull; yes, will build head from there and send you the new mileage report.<br /><br />(Glad your
weekendwith Irene is finally over - we had a few crises down here in New York; my fish were swimming in the back yard,
lotsof flooding, a few trees down - one of which only prevented from falling on the house by the 16K volt primary line.
Great!But we were luckier than many.)<br /><br />Lou Picciano<br /></div>
