答复: [RFR]8215623: Add incremental dump for jmap histo

[臧琳]

Dear Paul,

        Thanks for your review, I agree your suggestion to make incremental general. and

        I think the incremental file is better to be different with the file or the "out", because

in future the parallel histo will have each thread individually dump their data, and I want

it to be lock-free, so each thread need to dump to their own file. The final data

will be sum up and dump to the file that "filename=<file>" specified.  Moreover, the incremental

file contains intermediate data, if it is dumped to "<file>", it more or less "pollute" the final file (final file that contains lots intermediate data).

       so the logic is that incremental data will be dumped to a file named "Histo_Dump_Temp.dump",

if the "filename" is assigned, the "Histo_Dump_Temp.dump" will be generated under the same folder,

if "filename" not specified, it will dump to current dir. And if "chunkcount" is 0 or max_int, or "maxfilesize" is 0, the incremental dump will be disabled.

________________________________
·¢¼þÈË: Hohensee, Paul <hohensee at amazon.com>
·¢ËÍʱ¼ä: 2019Äê1ÔÂ1ÈÕ 4:56
ÊÕ¼þÈË: ê°ÁÕ; serviceability-dev at openjdk.java.net
Ö÷Ìâ: Re: [RFR]8215623: Add incremental dump for jmap histo


As for 8215622, update the copyright dates to 2019 please, since this won¡¯t get pushed until then.



You might generalize the implementation so that all inspections are done incrementally, and parameterize RecordInstanceClosure with the incremental threshold. ¡°incremental¡± could become ¡°chunkcount=<n>¡±, where <n> defaults to infinity (max value of size_t).



I¡¯d not use a default file name when ¡°chunkcount¡± is specified, I¡¯d just write to whatever the output stream is. ¡°chunkcount¡± is then independent of ¡°file¡±.



I¡¯d add another histo argument for the maximum file size, call it ¡°maxfilesize¡±, and parameterize RecordInstanceClosure with it. Default would be infinity (max value of size_t).



If you want to make it easy to use your 8k and 5mb chunkcount and maxfilesize combination, you could redefine your original ¡°incremental¡± argument as syntactic sugar for ¡°chunkcount=8k,maxfilesize=5m¡±.



INCREMENTAL_THRESHOLD and MAX_INCREMENTAL_FILESIZE become DEFAULT_CHUNKSIZE and MAX_FILE_SIZE, are both set to max size_t, and should be defined as ¡°const int¡± within RecordInstanceClosure.



Thanks,



Paul



From: serviceability-dev <serviceability-dev-bounces at openjdk.java.net> on behalf of ê°ÁÕ <zanglin5 at jd.com>
Date: Thursday, December 20, 2018 at 11:13 PM
To: "serviceability-dev at openjdk.java.net" <serviceability-dev at openjdk.java.net>
Subject: [RFR]8215623: Add incremental dump for jmap histo



Hi All,

       May I ask your help to review this patch for enhance jmap ¨Chisto.

It adds the ¡°incremental¡± arguments that allow jmap ¨Chisto to incrementally save the intermediate data into a temp file.

The intermediate data is dumped incrementally and write to a rolling file, which limit the size of the temp file to be small.

This is useful for user to get intermediate results when jmap/jvm process is killed accidentally. Especially when the heap is large.



This patch is also part of the enhancement described in https://bugs.openjdk.java.net/browse/JDK-8214535.



Webrev: http://cr.openjdk.java.net/~xiaofeya/8215623/webrev.00/

Bug: https://bugs.openjdk.java.net/browse/JDK-8215623



Thanks.



BRs,

Lin




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.java.net/pipermail/serviceability-dev/attachments/20190107/fc2776ad/attachment-0001.html>