Description of problem: RHQ server consumes 700Mb of resident memory on startup (compare to JON2.3.1, 400Mb). Clicking around various pages in the UI (without actually doing anything) raises the memory consumption to 960Mb in 15 clicks. (virtual memory = 1.7G). The automated tests do enough clicking to get OOM errors (heap space) Version-Release number of selected component (if applicable): b169eb6 How reproducible: Always, so far Steps to Reproduce: 1. Install RHQ 2. look at "top" or otherwise display memory usage 3. Click around the UI and watch it increase steadily Actual results: Expected results: Additional info:
java environment: [rhq@auto-rhq01 bin]$ java -version java version "1.6.0" OpenJDK Runtime Environment (build 1.6.0-b09) OpenJDK 64-Bit Server VM (build 1.6.0-b09, mixed mode)
Jeff, can you please elaborate more on "click around the UI" ? What parts did you especially try? Did you have many resources ?
Heiko, it doesn't matter what you click, every click raises the memory consumption 5-20mb, without exception. No i did not have many resources, just a single agent.
Sorry - to clarify, every click that causes a page load increases the memory usage. Obviously clicking into a text field wouldn't cause any memory increase on the server :)
I'm still seeing high memory consumption, and now the agent has also doubled in size. We really need to investigate this. The agent should not be taking 500MB. 4437 jonqa 25 0 1858m 971m 4824 S 0.0 48.5 80:05.26 java 519 root 18 0 1015m 526m 5652 S 0.0 26.3 4:51.48 java Build: f3cc7f7
Jeff, can you confirm whether or not you are still seeing this after the UI fixes we put in. I'm also interested if you are seeing excessive agent memory usage too.
I'm wondering if this could be related to https://bugzilla.redhat.com/show_bug.cgi?id=589674
[12:45] <cswiii> [12:40] <jweiss> last i checked, the memory usage was still rather high [12:45] <cswiii> [12:40] <jweiss> let me check last night's install [12:45] <cswiii> [12:41] <jweiss> 1678m virt, 717m resident [12:47] <ccrouch> i'd need to know what that looks like compared to 2.3.1 to know if thats an issue or not
FWIW, this is the results of an automation run against 2.3.1 that might be usable as some sort of baseline PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND [...] 1995 root 20 0 2082m 1.1g 13m S 0.3 54.9 8:17.82 java
QA Closing. We can reopen if this seems to show up again, but things look pretty comparable to me.
Mass-closure of verified bugs against JON.