RFR: 8294689: The SA transported_core.html file needs quite a bit of work [v2]
Poonam Bajaj
poonam at openjdk.org
Mon Oct 3 19:44:27 UTC 2022
On Mon, 3 Oct 2022 18:44:14 GMT, Chris Plummer <cjplummer at openjdk.org> wrote:
>> The main issues I set out to address are:
>>
>> - legacy Solaris references can be removed
>> - add help with dealing with Windows core files, most of important of which is use of PATH to find the dlls referenced by the core file.
>> - add help with dealing with macOS core files, most of important of which is using JAVA_HOME or DYLD_LIBRARY_PATH to find the dlls referenced by the core file.
>>
>> However, in the end it turned out to be a pretty substantial restructuring and rewrite.
>>
>> Here are links to the rendered new and old copies of the file:
>> [new file](https://htmlpreview.github.io/?https://github.com/openjdk/jdk/blob/6778a202a855a534754c4b47a6cd99cb912869d3/src/jdk.hotspot.agent/doc/transported_core.html)
>> [old file](https://htmlpreview.github.io/?https://github.com/openjdk/jdk/blob/master/src/jdk.hotspot.agent/doc/transported_core.html)
>
> Chris Plummer has updated the pull request incrementally with one additional commit since the last revision:
>
> some additional clarification on linux and macos
src/jdk.hotspot.agent/doc/transported_core.html line 14:
> 12: produced ("transported core dump"), debuggers (dbx, gdb, windbg or SA) do not
> 13: always successfully open the dump. This is due to kernel, library (shared
> 14: objects or DLLs) mismatch between core dump machine and debugger machine.
Looks like an additional space between _to_ and _library_.
src/jdk.hotspot.agent/doc/transported_core.html line 40:
> 38: Their pages are instead read from the executable and shared objects (or DLLs).
> 39: Therefore it is important to have a matching java executable and shared object
> 40: files on the debugger machine. The best way to guarantee this match is to match the
An extra space between _best_ and _way_.
src/jdk.hotspot.agent/doc/transported_core.html line 73:
> 71: If the debugger machine and core dump machine have identical Windows libraries, then you only
> 72: need to point SA to the location of the JDK bin directory. This is done by making sure
> 73: the bin directory is included in the <b>PATH</b> environment variable. If the windows
The location of jvm.dll (e.g. %Java_Home%\bin\server) should be included in the PATH environment variable. And if I remember correctly, the location of 'windbg' should also be added to PATH - this needs to be tested and confirmed though.
-------------
PR: https://git.openjdk.org/jdk/pull/10518
More information about the serviceability-dev
mailing list