Postgres on both nodes

Hi guys,
I work on NNMi 10.21, Red Hat 7.8, produced in application failover with NPS on another server.
I have the Postgress.OLD directory in /var/opt/OV/shared/nnm/databases, in both servers, both the active one and the standBy one. Does this directory need to exist on both servers? Or can it be deleted on the active node?

Thank you in advance

Antonio

Parents
  • Hi Antonio,

    There are two reasons why NNMi creates OLD subfolder. One is that you do an upgrade, if I remember well that was an action typical for a bigger version upgrade of Postgres, so 9.x, 10.x to 10.20. Also, I had a similar question from custom of NNMu under AF - changing the role of DB from hot_standby to standalone and backward also results in having it. My usual advice that is to not touch it, until you are in a shortage of diskspace...

    BR,

  • Hi Eugene,
    first of all thanks for your advice, yes I have 10:21 with patch 7, I should have installed 10 but soon Microfocus will upgrade me to the 2018 or 2020 version, so you advise me not to touch or delete the Postrgress directory. OLD?
    I have had space problems in the recent past, so I searched for documentation about it and found a document describing how to delete the directory and start from a clean situation. This is only in the Application Failover environment.
    But now I have cleaned up and I am under the warning thresholds regarding disk space, and I thought I would free up more space by deleting this Postgress.OLD, but I will follow your advice.
    Thanks

    Antonio

  • Verified Answer

    Hi Antonio,

    In case you are going to upgrade, and the Postgres version is going to change, I would do a full AF disabling process, not only to comment out the cluster name - this is my way. If you do so, you create a postgres trigger file, so the DB turns into a standalone instance, it should not have OLD folder at all, so you delete it when you need a space. Some upgrades need that extra space as 2 times of the current DB allocation, because it makes a flat and regular backup - so it creates an export folder, renames current DB to OLD, creates a new empty Postgres, and imports export to it.

    BR,

Reply
  • Verified Answer

    Hi Antonio,

    In case you are going to upgrade, and the Postgres version is going to change, I would do a full AF disabling process, not only to comment out the cluster name - this is my way. If you do so, you create a postgres trigger file, so the DB turns into a standalone instance, it should not have OLD folder at all, so you delete it when you need a space. Some upgrades need that extra space as 2 times of the current DB allocation, because it makes a flat and regular backup - so it creates an export folder, renames current DB to OLD, creates a new empty Postgres, and imports export to it.

    BR,

Children