but this is much much better.
Thanks for the help, and amazed at the responsiveness of Mr. Lane and this
community!
Tom Carroll
On Monday, December 10, 2018, 4:59:47 PM EST, Laurenz Albe
wrote:
Thomas Carroll wrote:
> We have a situation where a long-persistent Postgres connect
On Monday, December 10, 2018, 7:45:07 PM EST, Tom Lane
wrote:
Thomas Carroll writes:
> On Monday, December 10, 2018, 5:50:22 PM EST, Tom Lane
>wrote:
>> Also, as mentioned upthread, it'd be interesting to see if there's
>> a memory context dump s
On Monday, December 10, 2018, 5:50:22 PM EST, Tom Lane
wrote:
> Is the error message spelling really exactly "Cannot allocate memory"?
Yes - but I think the message is forwarded from Linux. Here is an example:
2018-12-08 00:00:00.070 EST,,,32506,,5bc71a25.7efa,25643,,2018-10-17 07:16:53
On Monday, December 10, 2018, 3:45:33 PM EST, Tom Lane
wrote:
Thomas Carroll writes:
> Postgres version: 10.5. work_mem setting: 4MB, shared_buffers setting: 800
> MB, connections typically around 30-40.
I imagine you checked this already, but ... what is temp_buffers set to
We have a situation where a long-persistent Postgres connection consumes more
and more memory. If ignored, we eventually get “Cannot allocate memory” errors
in the Postgres log. If still ignored, the box will eventually crash. This
takes about 3 weeks to happen. It issues a call to a single
