Re: Postgres stucks in deadlock detection - Mailing list pgsql-hackers

From Andres Freund
Subject Re: Postgres stucks in deadlock detection
Date
Msg-id 20180404185824.53vz6er5eqb24gau@alap3.anarazel.de
Whole thread Raw
In response to Re: Postgres stucks in deadlock detection  (Teodor Sigaev <teodor@sigaev.ru>)
List pgsql-hackers
Hi,

On 2018-04-04 21:55:10 +0300, Teodor Sigaev wrote:
> > > > Have you ever observed that in the field? This sounds more artificial
> > > > than real to me.
> > > 
> > > Zabbix storage with 50Gb WAL per hour on insertion.
> > 
> > That's not a very detailed description. Isn't that a largely insertion
> > only workload?
> 
> It's mostly insert load, collecting monitoring information. Also there are
> some read-only transactions, ~10 per second.

How are you running into deadlock detector issues then? Where do the
significant number of heavyweight lock weights come from?


> We didn't invent this from mind, client came to us with support request, now
> we have two different cases.

You gotta actually start describing those then.

Greetings,

Andres Freund


pgsql-hackers by date:

Previous
From: Pavan Deolasee
Date:
Subject: Re: pgsql: New files for MERGE
Next
From: Tom Lane
Date:
Subject: Re: pgsql: New files for MERGE