Re: BUG #14452: Unexpected transaction behaviour

From: John R Pierce <pierce(at)hogranch(dot)com>
To: pgsql-bugs(at)postgresql(dot)org, rjsdg(at)sapo(dot)pt
Subject: Re: BUG #14452: Unexpected transaction behaviour
Date: 2016-12-07 03:56:16
Message-ID: 7ff1a524-c292-eda3-6bd3-7d3c5976b419@hogranch.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

On 12/6/2016 6:46 PM, rjsdg(at)sapo(dot)pt wrote:
> The system normal throughput is about 110 transactions/min constantly
> without degradation. Recently a new Java module of the application was
> installed and when the app startups the system throughput starts to degrade
> to the point that reaches 50 transactions/min, it’s not an immediate
> downfall in the throughput but it looses like 5 to 10 transactions per
> minute until it reaches 50 transactions per minute. At this point if we
> terminate the only existing backend pid of the new Java module (that is in
> “idle in transaction” with a backend_xmin value) immediatly in the next
> minute the throughput rises again to 110 transactions/min.

'idle in transaction' for more than a short amount of time is bad news,
it blocks all maintenance activity like vacuuming, and causes frequently
updated tables and indexes to bloat badly.

--
john r pierce, recycling bits in santa cruz

In response to

Browse pgsql-bugs by date

  From Date Subject
Next Message akshay.ballarpure 2016-12-07 11:06:18 BUG #14453: Growing file descriptors on IDLE connections
Previous Message rjsdg 2016-12-07 02:46:25 BUG #14452: Unexpected transaction behaviour