Re: XID wraparound and busy databases - Mailing list pgsql-hackers

From Tom Lane
Subject Re: XID wraparound and busy databases
Date
Msg-id 7058.1187196592@sss.pgh.pa.us
Whole thread Raw
In response to Re: XID wraparound and busy databases  (Heikki Linnakangas <heikki@enterprisedb.com>)
Responses Re: XID wraparound and busy databases  (Decibel! <decibel@decibel.org>)
Re: XID wraparound and busy databases  (Darcy Buskermolen <darcy@dbitech.ca>)
Re: XID wraparound and busy databases  (Zoltan Boszormenyi <zb@cybertec.at>)
Re: XID wraparound and busy databases  ("Simon Riggs" <simon@2ndquadrant.com>)
List pgsql-hackers
Heikki Linnakangas <heikki@enterprisedb.com> writes:
> Maybe we can do something to reduce the xid consumption? For example,
> reuse xids for read-only queries.

Hmm, that's an idea.

More simply, just keep the current transaction open (resetting
everything but the XID) if we have made no changes by the time we're
told to commit or rollback ... which is something we track already,
so as not to waste cycles on useless commit XLOG records.

You'd want some upper limit on transaction lifetime, so as to avoid the
"long lived transactions hurt VACUUM" problem, but even reusing a single
xact for a few seconds would pretty much eliminate this issue, I bet.
It's hard to see how anyone could be doing 6K xacts/sec unless most are
read-only.
        regards, tom lane


pgsql-hackers by date:

Previous
From: "Dawid Kuroczko"
Date:
Subject: Index Tuple Compression Approach?
Next
From: Decibel!
Date:
Subject: Re: XID wraparound and busy databases