Zum Inhalt springen
Sie müssen registriert sein, um mit der Community zu interagieren.
Alle Beiträge Personen Abzeichen
Stichwörter (Alle anzeigen)
odoo accounting v14 pos v15
Über dieses Forum
Diese Frage wurde gekennzeichnet
performancehardwaresizingworkers
2 Antworten
13396 Ansichten
Avatar
Martin Schubert

In 10 days our product is featured in a tv-show at prime time.

We expect 50k visitors on our website.

Cause we are a small company (5 users) we would like to avoid a load balancing server scenario for this single visitor peak.

Question 1: Is this configuration working for this case?

- Hardware: Google Cloud instance with 64 vCPUs, 240gb RAM, 20gb SSD

- Debian: default

- Nginx: default

- Odoo: 10 community - fresh database with 30 products and this modules: Ecommerce, Sales, Purchase, Manufacturing, Inventory, Accounting, no custom modules

workers = 100

limit_memory_hard = 1677721600 (1,6GB)

limit_memory_soft = 1073741824 (1GB)

limit_request = 8192

limit_time_cpu = 60

limit_time_real = 120

- Postgres: max_connections = 6500 (100 worker * 64 connections)

Question 2: How can I simulate the load to locate bottelnecks?

Has anyone experience with www.loader.io ?

Avatar
Verwerfen
Avatar
Martin Schubert
Autor Beste Antwort

TV-Show is done and Odoo ran stable.

Visitors: 27k, concurrend 10k

with this settings:

- Hardware: Google Cloud instance with 64 vCPUs, 240gb RAM, 20gb SSD

- workers = 100

- limit_memory_hard = 1677721600 (1,6GB)

- limit_memory_soft = 1073741824 (1GB)

- http://pgtune.leopard.in.ua/

Load average:

CPUs: 30%

RAM: 25GB

Avatar
Verwerfen
Avatar
Dipak Shah
Beste Antwort

Hello Martin Schubert,

By Setting only workers is not enough you have to configure few things,

: (parameter of odoo/postgres database).

1> IN odoo config :   

    limit-memory-soft,limit-memory-hard

2> IN postgres (database) :

    max_connections,shared_buffers,effective_cache_size,work_mem

    maintenance_work_mem,checkpoint_segments,checkpoint_completion_target

    wal_buffers,default_statistics_target.

  • To share the workload between workers mostly you need to divide and allocate memory in such a way so its depending on the system config. (e.g Hard disc Type, Memory (RAM), Size of it,version of postgres and your O.S) .

You do not need to worry about it ,just go through either of below URL to boost up your server speed.

> http://pgtune.leopard.in.ua

> https://github.com/gregs1104/pgtune

HOPE this will help you a lot !

Many Thanks.


Avatar
Verwerfen
Verknüpfte Beiträge Antworten Ansichten Aktivität
Odoo.sh performance / sizing
performance sizing odoo.sh
Avatar
Avatar
Avatar
Avatar
4
Mai 22
13593
Do Database Workers affect the Odoo SH platform performance?
performance slow workers OdooSH
Avatar
Avatar
1
Aug. 23
5226
Seeking Odoo Server Performance configuration/ advice / tips / benchmarks for 60 concurrent users
performance postgresql workers tuning odoo
Avatar
0
März 24
6570
Odoo 10 CE performance tuning - enabling workers Gelöst
performance serversecurity workers odoo v10.0
Avatar
Avatar
1
Aug. 22
31652
Odoo 8 with 10 users tips to fine tuning the database and workers
configuration performance postgresql9.3 workers tuning
Avatar
0
Mai 16
11551