RDS malfunctioning after upgrading 9.6 to 12.7 (accidentally in the same day that we had AWS outage)

Question

We do not have a technical support account however I am going to ask about the issue that we have lately on IO in our RDS access(after AWS outage). We were upgrading RDSs from 9.6 to 12.7 (then to 13.3 due to the EOL for 9.6 on Jan 18th 2022) and it was right at the same time of the outage. After the upgrade to 12.7, we had many issues(it got a little bit better on 13.3). We still have issues and our application got affected by a malfunctioning and slowness in the RDSs. I am not sure if we still have some technical issue that causes the slowness and IO issues on our RDSs or if that is another independent issue. That is why I wanted to make sure everything works fine on AWS first and then try to troubleshoot on our side.

Answer

You may have already know this and have done this, but in case you have not...  PostgreSQL does not carry over statistics between major versions.  So, when you do an upgrade from 9.6 to 12 to 13, the statistics used by the query optimizer are lost.  This can lead to poor performance and increase in IO.  The takeaway is be sure to do ANALYZE after a major upgrade.  Here is a blog post that talks about this: https://medium.com/aubergine-solutions/trust-me-you-dont-want-to-miss-this-postgresql-command-after-the-upgrade-99200cbfabaf

RDS malfunctioning after upgrading 9.6 to 12.7 (accidentally in the same day that we had AWS outage)

관련 콘텐츠