Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
HBase >> mail # user >> GC and High CPU


Copy link to this message
-
Re: GC and High CPU
so far its going well, no more crazy GC, while load remains the same.
2011-05-18T11:01:53.149-0700: 52288.388: [CMS-concurrent-preclean:
5.764/216.720 secs] [Times: user=103.71 sys=15.85, real=216.68 secs]
2011-05-18T11:01:53.149-0700: 52288.388:
[CMS-concurrent-abortable-preclean-start]
2011-05-18T11:01:53.270-0700: 52288.509:
[CMS-concurrent-abortable-preclean: 0.117/0.121 secs] [Times:
user=0.13 sys=0.01, real=0.13 secs]
2011-05-18T11:01:53.271-0700: 52288.510: [GC[YG occupancy: 104181 K
(191744 K)]52288.510: [Rescan (parallel) , 0.0217820 secs]52288.532:
[weak refs processing, 0.0010760 secs] [1 CMS-remark:
11810687K(12075008K)] 11914869K(12266752K), 0.0230420 secs] [Times:
user=0.19 sys=0.00, real=0.02 secs]
2011-05-18T11:01:53.294-0700: 52288.534: [CMS-concurrent-sweep-start]

preclean took a long time, but other then that, we remain super
stable, and using HEAP to the max.

-Jack
On Mon, May 16, 2011 at 7:41 PM, Andrew Purtell <[EMAIL PROTECTED]> wrote:
> This is interesting because our "conventional wisdom" is those settings should increase the chance of stop-the-world GC and should be avoided.
>
>  - Andy
>    (who always gets nervous when we start talking about GC black magic)
>
>> From: Jack Levin <[EMAIL PROTECTED]>
>> Subject: Re: GC and High CPU
>> To: [EMAIL PROTECTED]
>> Date: Monday, May 16, 2011, 5:06 PM
>>
>> Those are the lines I added:
>>
>> -XX:+CMSIncrementalMode \  <--------
>> -XX:+CMSIncrementalPacing \ <-------
>> -XX:-TraceClassUnloading <------
>>
>> -Jack
>> (used CMS before)
>>
>> On Mon, May 16, 2011 at 4:19 PM, Stack <[EMAIL PROTECTED]>
>> wrote:
>> > So, the change is that you started using CMS?  You
>> > were using the default GC previous?
>> >
>> > ParNew is much bigger now.
>
>