RE: Memory pressure on Cloud at customer solution with EXADATA
Date: Mon, 20 May 2019 19:33:54 -0700
Message-ID: <005201d50f7d$a9ae51c0$fd0af540$_at_comcast.net>
If you are getting OOM messages then OOM killer could have kicked in and the VM will crash. This evening happens on Physical servers crashing the physical server OS.
From: oracle-l-bounce_at_freelists.org <oracle-l-bounce_at_freelists.org> On Behalf Of Andrew Kerber
Sent: Monday, May 20, 2019 7:28 PM
To: jack_at_vanzanen.com
Cc: Chris Taylor <christopherdtaylor1994_at_gmail.com>; ORACLE-L <oracle-l_at_freelists.org>
Subject: Re: Memory pressure on Cloud at customer solution with EXADATA
Normally the oracle alert logs will say something about swapping if it’s happening.
Sent from my iPad
On May 20, 2019, at 19:55, Jack van Zanen <jack_at_vanzanen.com <mailto:jack_at_vanzanen.com> > wrote:
Jack van Zanen
This e-mail and any attachments may contain confidential material for the sole use of the intended recipient. If you are not the intended recipient, please be aware that any disclosure, copying, distribution or use of this e-mail or any attachment is prohibited. If you have received this e-mail in error, please contact the sender and delete all copies. Thank you for your cooperation
On Tue, May 21, 2019 at 10:28 AM Chris Taylor <christopherdtaylor1994_at_gmail.com <mailto:christopherdtaylor1994_at_gmail.com> > wrote:
You should be seeing OOM messages and killed processes before a node crash i would think. Did the whole vm crash, or did the db crash only? If I understood you correctly, the whole vm crashed.
That seems like an odd scenario for an OOM condition - especially with 200GB left for OS and processes.
Chris
On Mon, May 20, 2019, 6:54 PM Jack van Zanen <jack_at_vanzanen.com <mailto:jack_at_vanzanen.com> > wrote:
Hi All,
Oracle 12.2.0.1
PSU Jan 2019
Kernel version: 4.1.12-124.24.3.el6uek.x86_64 #2 SMP Mon Jan 14 15:08:09 PST 2019 x86_64
we have a 2 node RAC/Exadata solution and have configured about 500GB of 720GB memory for Huge Pages.
Now this is over configured at the moment as we have decommissioned some databases. This is our combined TEST/DR environment and we could spin up more containers at any time.
Last week we had node 1 crash and oracle support came back saying it was due to memory pressure and we need to configure fewer Huge Pages.
While in theory that makes sense but I am struggling to understand why a system with 200G+ memory not allocated to huge pages is crashing due to memory issues. I have checked our alert logs and the databases start using huge pages.
Anyone here can explain to me how this can happen, as oracle support keep repeating to lower the configured huge pages.
Jack van Zanen
This e-mail and any attachments may contain confidential material for the sole use of the intended recipient. If you are not the intended recipient, please be aware that any disclosure, copying, distribution or use of this e-mail or any attachment is prohibited. If you have received this e-mail in error, please contact the sender and delete all copies. Thank you for your cooperation
-- http://www.freelists.org/webpage/oracle-lReceived on Tue May 21 2019 - 04:33:54 CEST