The problem of JIRA Memory

[yunfengg]

Our JIRA is facing high response time issue, and when i look into the system info, i found the the memory is going to be used up. But we set 8G to JVM before and the issue has no been occured.

In addition, i found the PS Old Gen occupied 100%. So i need to know what's the root cause and how do we configure the memory of JIRA

BTW,here is the info of our JIRA:

Issues	129694
Projects	496
Custom Fields	110
Workflows	27
Attachments	77885
Comments	261754
Users	7112
Groups	25

[Rahul Aich [Nagra]]

Hi Yunfeng

Can I ask you if you have got 32 bit or a 64 bit jira installation. And which version of jira you are on.

Note that on a 32 bit installation there is a capping (around 1.25GB) on the amount of memory that you can allocate to jira. If you are on a 32 bit then i recommend upgrading your jira installation to 64 bit

Rahul

[Nic Brough -Adaptavist-]

Simply throwing more memory at a JVM is not a good answer. You need to look at what the problem actually is.

Start with https://confluence.atlassian.com/display/JIRA/Increasing+JIRA+Memory

[yunfengg]

I wonder why Code Cache and PS Old Gen occupies about 98% now

[Nic Brough -Adaptavist-]

Because you're not giving it enough room in your memory settings. Try changing them as per the doc I linked to.

[yunfengg]

@Nic, Can you give me some suggestion about the memory configuration both for JIRA and DB based on my JIRA enviroment(We deployed it on a Linux server(128G memory) and DB is oracle and is on a VM (128G memory), the detail usage information as you can see below)

[Zul NS [Atlassian]]

I think that giving 8GB for the instance is too much, might want to lower it down to around 2GB. What is your previous Xmx settings before you raising it to 8GB? I would suggest to increase the Heap space of 256MB slowly up.

[Nic Brough -Adaptavist-]

As well as going back to a smaller overall memory space and then building up slowly as Zulfadli says, you need to tune the rest of the settings in line with each increase. It really is not as simple as just throwing more memory at it, there are consequences. You need to read up on Java memory tuning.