Check your Odoo setup in 3 stages
Check your Odoo implementation
If your implementation was done correctly, your system should not only run very fast but also be very stable. If this is not the case, it is great time to check your Odoo implementation. This means the technical deployment of either Odoo and / or PostGreSQL was not done properly. Before choosing a partner, have you followed the “BE SAFE” recommendations?
If you have not, you already ahving bugs, error messages, server restarts et…. etc… This is the common lot of unexperienced implementations. Now it is time to cross-check your implementation. You can easily do so by checking the following points:
Phase 1: The Odoo configuration file
Obtain a copy of the odoo configuration file “.conf”.
It the line “workers” is absent from the file OR the value is “workers = 0“, then the multi-worker mode is not activated. System will be slow, buggy and may eventually crash underheavy use.
Phase 2: PostgreSQL setup
Obtain a copy of the PostgreSQL confifuration file: postgresql.conf. look for the following criteria.
If you find these lines, then setup was not done properly:
#default_statistics_target = 100
#constraint_exclusion = partition
#checkpoint_completion_target = 0.5
#checkpoint_segments = 3
#max_connections = 100
#maintenance_work_mem = 16MB
#effective_cache_size = 128MB
#work_mem = 1MB
#wal_buffers = -1
#shared_buffers = 128MB
These are the standard install defaults. The hash tag in front meeans that all these parameters were commented out and therefore NO setup was done. PostgreSQL runs un minimal mode, slowly and without using your system resources. They DO NOT WORK properly until the correct values have been calculated and input according to your system specs and users… Read more about this here…
Phase 3: Install “pg_top”
pg_top will give you a list of the active connections on your system. If the list is longer than the number of current users, then you are experiencing memory leakage. AGAIN, this means your system will crash soon when the total amount of memory has been used and is saturated with unused connections. Read more about this here…
Again, if none or only a few of these key points have not been setup properly, then it is very likely the system is either buggy, unstable or slow. It is mandatory to fix these issues before moving to production.