-
Type:
Bug
-
Resolution: Low Engagement
-
Priority:
Low
-
None
-
Affects Version/s: 7.0.0, 7.1.0, 7.2.0, 7.3.0, 7.4.0, 7.5.0
-
Component/s: Documentation - All
-
None
-
7
-
1
-
Severity 3 - Minor
-
0
Customer raised a community issue in regards to this page:
https://confluence.atlassian.com/adminjiraserver075/jira-applications-installation-requirements-935390824.html
The focus here is the Server side installation requirements for production section which reads:
- For a small number of projects (less or equal to 100) with 1,000 to 5,000 issues in total and about 100-200 users, a recent server (multicore CPU) with 2GB of available RAM and a reasonably fast hard drive (7200 rpm or faster) should cater for your needs.
- For more than 100 projects you should monitor JIRA memory usage and allocate more memory if required.
- If your system will experience a large number of concurrent requests, running JIRA applications on a multicore CPU machine will increase the concurrency of processing the requests, and therefore, speed up the response time for your users.
- For reference, we have a server that has a 2 Intel(R) Xeon(R) CPU E5520 @ 2.27GHz (16 logical cores) with 32GB of RAM. This server runs Apache, various monitoring systems, and two JIRA application instances:
- Our public site has approximately: 145,000 issues, 255,000 comments, 120 custom fields, and 115 projects.
- Our support site has approximately: 285,000 issues, 2,500,000 comments, 75 custom fields, and 22 projects.
The values and numbers mentioned here are wildly off what we use in regards to Sizing estimates for customers such as in JIRA Sizing Guide - Atlassian Documentation.
The major points I'd like to see addressed are
- the phrase "with 2GB of available RAM". The Server itself should just about always have 8GB of dedicated system memory, and Jira's JVM should typically have at least 1GB of xmx assigned to it, even for the smallest installations.
- Creating 100 projects creates a lot more overhead than this guide implies. That's because each created project can create new workflows, new custom fields, new permissions schemes, new screens, etc.
- Also the information in regards to JAC and SAC/GSAC is no longer current and accurate. I know GSAC is using datacenter deployment now. I understand if this change might require a separate ticket, but it would be great if we can keep these updated to reflect our current use.