JEP 271: Unified GC Logging - First pre-review

[Bengt Rutisson]

On 2015-10-27 15:39, Kirk Pepperdine wrote:
>
>> On Oct 26, 2015, at 11:22 PM, Bengt Rutisson 
>> <bengt.rutisson at oracle.com <mailto:bengt.rutisson at oracle.com>> wrote:
>>
>>
>> Hi Poonam,
>>
>> Thanks for looking at this!
>>
>>
>> On 2015-10-26 21:22, Poonam Bajaj Parhar wrote:
>>> Hello Bengt,
>>>
>>> I am going to try out the new logging myself but some quick feedback 
>>> from the examples output you had provided:
>>>
>>> *-Xlog:gc:*
>>>
>>> 1. It is great to have the start and end timestamps included on the 
>>> same GC log line.
>>
>> Great! :)
>>
>>>
>>> 2.
>>> /[0.643s][info ][gc] GC#5 GC young (G1 Evacuation Pause) end 
>>> 91M->61M(128M) (0.632s, 0.643s) 11.399ms////[1.499s][info ][gc] GC#16 GC remark end 73M->73M(128M) (1.486s, 
>>> 1.499s) 12.545ms [1.500s][info ][gc] GC#16 GC cleanup end 
>>> 73M->73M(128M) (1.499s, 1.500s) 0.729ms/
>>>
>>> It does not make much of a sense here to include 'end' in the log 
>>> lines. If there is no line with a 'start' then I think it does not 
>>> serve any purpose to have 'end' for these GC lines.
>>
>> I see your point. However, when the start logging is included I think 
>> it makes sense to have some marker on this line saying that it is the 
>> end logging. I would prefer that the line looks exactly the same in 
>> both cases. But I am very open to suggestions for another way to 
>> format the line. I agree that it looks odd with "end" when there is 
>> no start. On the other hand the logging is done at the end of the GC, 
>> which could be relevant to include in the information. I've been 
>> thinking about "done" instead. Or "completed". Any other suggestions?
>
> How about removing start end. It’s quite obvious in the log were the 
> record starts and ends without these extra decorations.

Yes, I think that is the best suggestion so far. I'll do that.

Bengt

>
> Kind regards,
> Kirk
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/hotspot-gc-dev/attachments/20151028/78ca51ce/attachment.htm>