Mailing Lists: Apple Mailing Lists

Image of Mac OS face in stamp
 
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Help!: Kernel Panics under heavy socket connections, Java 1.5 causing them?



Hi Ben,

Would it be possible for us to get the stress tests you mention? Along with some information on how to run them. You can send them to me offline.

Thanks,
Pratik

On 4/17/07 3:55 PM, Ben Spink wrote:
I have kernel panics a lot lately. I have yet to determine the root cause, but I have my theories and guesses. I will explain the best I can, but I think it has to do with the current Java 1.5 version on OS X 10.4.9.

My test machine:

Mac Mini Core Duo 1.66, 2GB RAM, various HD's USB and Firewire, OS 10.4.9, Java 1.5

Since around march or so I have noticed Kernel panics occasionally. My first guess was RAM. I replaced the RAM twice now and currently have Crucial DIMM's in place...so I am fairly trusting in them. (I have dealt with bad ram before...these panics don't act like panics from bad ram.) Especially considering I can make the Kernel panic happen almost on demand. And its not based on the RAM I'm using.

No new hardware has been added since the machine was stable to what its like now.

During some load testing on my server, CrushFTP4, when I would hammer my server with thousands of connections as fast as I could create them (100 threads just creating new Socket("192.168.1.2",8080);), after a few hundred my machine would kernel panic. Temperature was around 57C. CPU was maxed at 100% on both cores. Only other apps running were Jbuilder 7, Mail, and TextWrangler.

This test was a couple months ago.

A couple days ago, I was under a DOS attack and was getting around 100 connections a second which my server had banned and was disposing of them as quickly as they were coming in. My CPU was sitting around 10% or so on both cores dealing with this. However, they were able to kernel panic me after just a few minutes of this. 3 times in one day before they gave up. Further, once they did this, a weird applescript error message would appear in the Finder...my code definitely isn't doing any applescript at all, so I don't know where this was coming from.

This was what finally made me think it had to do with Java as it happened so many times in a row. If I didn't have my server running, I was fine. As soon as I started it, within a minute or two a panic would occur.

Other than this, I never have a panic....ever.

More incidents...a client of mine just got a new 8 core tower with 4GB of RAM, OS 10.4.9, and Java 1.5. He was reporting panics while using CrushFTP as well. Not consistent. initially I thought there was something wrong with his machine, but after a few days of use, he said the only panics he got was when using CrushFTP....a Java app.

I then switched his machine to Java 1.4.2, and my machine back to Java 1.4.2. Since then he has not reported another kernel panic to me, and neither have I had another panic. Normally I could almost count on a panic every night.

Has anyone else noticed this? I've done all the standard things, repair permissions, reinstall combo update for 10.4.9, etc. But considering a brand new virgin machine from Apple was having the same issues, I don't think that it is something specific to my machine. And my code isn't leaking threads or something like that. Besides the same stress test on an XP machine did not crash the XP machine...it just chugged right along as I would have expected.

Here is the last panic I had (they all look almost identical to this):

Sun Apr 15 07:34:55 2007
panic(cpu 1 caller 0x0019A566): interlock timeout for mutex 0x4bdd99c
Backtrace, Format - Frame : Return Address (4 potential args on stack)
0x255f3db8 : 0x128d08 (0x3cb134 0x255f3ddc 0x131de5 0x0)
0x255f3df8 : 0x19a566 (0x3d0020 0x4bdd99c 0x255f3e38 0x4bdc838)
0x255f3e28 : 0x197e97 (0x4bdd99c 0x206 0x255f3ec8 0x331d2f)
0x255f3e38 : 0x331d2f (0x4bdd99c 0x4a371f4 0x319f 0x0)
0x255f3ec8 : 0x3335f4 (0x500c7d0 0x4bdd968 0x4a371f4 0x199193)
0x255f3f18 : 0x3336be (0x4a371f4 0x4f5 0x500c7d0 0x3)
0x255f3f58 : 0x379e23 (0x4a371f4 0x48f7e10 0x48f7e54 0x0)
0x255f3fc8 : 0x19b17e (0x476e848 0x19b160 0x4 0x207) No mapping exists for frame pointer
Backtrace terminated-invalid frame pointer 0xb0a04908


Kernel version:
Darwin Kernel Version 8.9.1: Thu Feb 22 20:55:00 PST 2007; root:xnu-792.18.15~1/RELEASE_I386



*********


I'm open to suggestions...

Things I am still going to try...re running my stress test on 1.4.2, then again on 1.5. Then again on 1.6 (that is available for download now isn't it?).

Thanks,
Ben
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Java-dev mailing list      (email@hidden)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/java-dev/email@hidden

This email sent to email@hidden

_______________________________________________ Do not post admin requests to the list. They will be ignored. Java-dev mailing list (email@hidden) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/java-dev/email@hidden

This email sent to email@hidden
References: 
 >Help!: Kernel Panics under heavy socket connections, Java 1.5 causing them? (From: Ben Spink <email@hidden>)



Visit the Apple Store online or at retail locations.
1-800-MY-APPLE

Contact Apple | Terms of Use | Privacy Policy

Copyright © 2007 Apple Inc. All rights reserved.