Re: systemd issues

  • From: "Radoulov, Dimitre" <cichomitiko@xxxxxxxxx>
  • To: Michael Schmitt <mschmitt@xxxxxxxxxxxx>, "oracle-l@xxxxxxxxxxxxx" <oracle-l@xxxxxxxxxxxxx>
  • Date: Thu, 2 May 2019 23:49:05 +0200

Hi Michael,

yes, I get the same behavior when I don't use systemctl to stop/start the service. I'll let you know if I find how to deal with this scenario.

Anyway, I think that you could avoid the problem if you start the database using the systemctl command after the patching.


Regards
Dimitre


On 02/05/2019 19.40, Michael Schmitt wrote:


Thanks,

Yes, it is really the service definition that we are struggling with. We have found when we use type=forking , systemd will kill off our oracle processes during the next server shutdown instead of waiting for the database(s) to shutdown.  We are trying to avoid a db crash as part of a server shutdown.  It works fine if we do not bring down the database down between server boots.  This is even with killmode=none.  If you haven’t run into that problem, we might need to test it again just incase the test got messed up for some other reason

Thanks

*From:*oracle-l-bounce@xxxxxxxxxxxxx <oracle-l-bounce@xxxxxxxxxxxxx> *On Behalf Of *Radoulov, Dimitre
*Sent:* Thursday, May 02, 2019 2:52 AM
*To:* Michael Schmitt <mschmitt@xxxxxxxxxxxx>; oracle-l@xxxxxxxxxxxxx
*Subject:* Re: systemd issues

Hello,

I wrote a simple script that seems to work, it starts/stops multiple Oracle instances and listeners (it looks for entries with Y in oratab).
The content of the systemd service definition is reported in the header.
You could try to use it as a template and adapt it for your environment.

Dimitre

On 01/05/2019 22.00, Michael Schmitt wrote:

    Hi Oracle-l

    Has anyone had any luck getting systemd to play nicely with Oracle
    databases?  It seems like every document I have read offers
    different suggestions and each of those come with their own set of
    issues

    The main issue we are having is systemd killing the oracle
    processes in different circumstances (when databases are manually
    restarted due to cold backups/patching/etc).  We really just want
    systemd to call our custom startup/shutdown scripts when the
    server reboots (and shutdown the database in advance of the
    network), instead of trying to manage the database processes

    Thanks in advance

Other related posts: