RFR: 8306437: jhsdb cannot resolve image/symbol paths being used for analysis of Windows coredumps
Chris Plummer
cjplummer at openjdk.org
Wed Apr 19 18:42:43 UTC 2023
On Wed, 19 Apr 2023 10:39:40 GMT, Alexey Pavlyutkin <duke at openjdk.org> wrote:
> Hi! The patch fixes image/symbol lookup by jhsdb on alanysis Windows coredump. It uses executableName as a hint prepending image path with
>
> `<executable_folder>;<executable_folder>\server`
>
> and symbol path with
>
> `srv*https://msdl.microsoft.com/download/symbols;<executable_folder>;<executable_folder>\server`
>
> where the first bit points to Windows symbols located on remote server
Did you read the "Using transported core dumps on Windows" section of transported_core.html?
https://github.com/openjdk/jdk/blob/master/src/jdk.hotspot.agent/doc/transported_core.html#L70
I'd like to know if you tried everything mentioned and still had issues. I reworked this section recently, but didn't have any first hand experience to go with. Just the old text and some input from those were using jhsdb with Windows.
It would be good if you could update this section to reflect your changes, and also add something regarding `sun.jvm.hotspot.debugger.windbg.imagePath` and `sun.jvm.hotspot.debugger.windbg.imagePath`, which I wasn't aware of when I recently updated this doc.
I understand your changes regarding looking relative to the executable name, but I don't understand how the following works. These are Windows services I'm not familiar with. Can you explain?
// request Windows symbol table from remote server...
symbols.append("srv*https://msdl.microsoft.com/download/symbols;");
-------------
PR Comment: https://git.openjdk.org/jdk/pull/13530#issuecomment-1515198314
More information about the serviceability-dev
mailing list