回复: 回复: 回复: 回复: JVM crashes constantly when High GC happens
shang xinli
shangxinli at hotmail.com
Tue Apr 23 17:55:20 UTC 2019
Yes, this happens on multi machines but only on those with a specific disk driver. I am thinking the device drivers which runs in kernel mode could corrupt other's memory. The smartctl is kicked off by command line which makes system call by issuing ioctl like what I copy & pasted in last email.
No, we don't swap to disk.
There is one question I would need help before I starting investigation. When a device driver servicing one user mode application's system call(ioctl), can that device driver corrupt other user mode application's memory(heap in our case)? How?
发送自 Outlook<http://aka.ms/weboutlook>
________________________________
发件人: Andrew Haley <aph at redhat.com>
发送时间: 2019年4月23日 10:03
收件人: shang xinli; Kim Barrett; shade at redhat.com
抄送: hotspot-gc-dev at openjdk.java.net
主题: Re: 回复: 回复: 回复: JVM crashes constantly when High GC happens
On 4/23/19 3:16 PM, shang xinli wrote:
> Thanks, Andrew! I agree it could be because of memory corruption.
> The issue only reproduced when a tool smartctl is kicked off.
More information needed.
Does this happen on moe than one machine?
How is smartctl kicked off? What commands does it use? Are you
swapping onto this disk?
> It has system call as below. Do you think it is possible that the
> device driver corrupts JVM's memory? How do I start investigating?
> Any use tool?
It's going to be extremely difficult. I'm thinking about kernel bugs. Is
your kernel up to date? All the device drivers?
--
Andrew Haley
Java Platform Lead Engineer
Red Hat UK Ltd. <https://www.redhat.com>
EAC8 43EB D3EF DB98 CC77 2FAD A5CD 6035 332F A671
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/hotspot-gc-dev/attachments/20190423/76ef7050/attachment.htm>
More information about the hotspot-gc-dev
mailing list