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

Switch to Plain View
HBase >> mail # user >> Amazon EC2 Virtualization vs. Dedicated Servers and Garbage Collection Times


+
Fuad Efendi 2011-08-08, 19:38
+
Jean-Daniel Cryans 2011-08-29, 18:39
Copy link to this message
-
Re: Amazon EC2 Virtualization vs. Dedicated Servers and Garbage Collection Times
Then how to explain this unpredictable 629 seconds with completely
underloaded server:
Total time for which application threads were stopped: 0.0004060 seconds
Total time for which application threads were stopped: 629.5515080 seconds
(!)
Total time for which application threads were stopped: 0.0006960 seconds
Total time for which application threads were stopped: 0.0001590 seconds
Total time for which application threads were stopped: 0.0003250 seconds

And this is still unclear too:
[Times: user=0.00 sys=0.00, real=158.65 secs]
What is "real" in EC2 terms? Broken router?!


On 11-08-29 2:39 PM, "Jean-Daniel Cryans" <[EMAIL PROTECTED]> wrote:

>Looks normal to me considering the platform. It's not so much a GC as
>the machine was unavailable for more than 2 minutes since there's no
>user or sys CPU involved.
>
>J-D
>
>On Mon, Aug 8, 2011 at 12:27 PM, Fuad Efendi <[EMAIL PROTECTED]> wrote:
>> Hello,
>>
>>
>> How to explain that:
>>
>> 2011-08-08T19:02:24.947+0000: 14564.360: [GC 14564.360: [ParNew:
>> 52681K->6528K(59008K), 158.6620830 secs] 2285054K->2240073K(4187776K)
>> icms_dc=0 , 158.6622690 secs] [Times: user=0.00 sys=0.00, real=158.65
>>secs]
>> Total time for which application threads were stopped: 158.6632510
>>seconds
>>
>>
>> QUESTION: How can? Zero, Zero, and real=158 seconds?!!
>>
>>
>> However, in most cases (99.99%) I have
>> 2011-08-08T18:55:20.811+0000: 14140.225: [GC 14140.225: [ParNew:
>> 54381K->2110K(59008K), 0.0286450 secs] 1651815K->1599544K(4187776K)
>> icms_dc=0 , 0.0287720 secs] [Times: user=0.00 sys=0.00, real=0.03 secs]
>> Total time for which application threads were stopped: 0.0296330 seconds
>> 2011-08-08T18:55:21.181+0000: 14140.594: [GC 14140.594: [ParNew:
>> 54588K->3056K(59008K), 0.0388800 secs] 1652022K->1600490K(4187776K)
>> icms_dc=0 , 0.0390230 secs] [Times: user=0.00 sys=0.00, real=0.04 secs]
>> Total time for which application threads were stopped: 0.0398510 seconds
>> 2011-08-08T18:55:21.527+0000: 14140.940: [GC 14140.940: [ParNew:
>> 55536K->4812K(59008K), 0.0326740 secs] 1652970K->1602247K(4187776K)
>> icms_dc=0 , 0.0328500 secs] [Times: user=0.00 sys=0.00, real=0.04 secs]
>> Total time for which application threads were stopped: 0.0338200 seconds
>> 2011-08-08T18:55:21.868+0000: 14141.281: [GC 14141.281: [ParNew:
>> 57258K->5232K(59008K), 0.0091650 secs] 1654693K->1604268K(4187776K)
>> icms_dc=0 , 0.0093210 secs] [Times: user=0.00 sys=0.00, real=0.01 secs]
>> Total time for which application threads were stopped: 0.0101470 seconds
>> 2011-08-08T18:55:22.179+0000: 14141.592: [GC 14141.592: [ParNew:
>> 57712K->6521K(59008K), 0.0307120 secs] 1656748K->1608576K(4187776K)
>> icms_dc=0 , 0.0308700 secs] [Times: user=0.00 sys=0.00, real=0.03 secs]
>> Total time for which application threads were stopped: 0.0318530 seconds
>> 2011-08-08T18:55:22.603+0000: 14142.016: [GC 14142.016: [ParNew:
>> 59001K->364K(59008K), 0.0471390 secs] 1661056K->1608313K(4187776K)
>>icms_dc=0
>> , 0.0472720 secs] [Times: user=0.00 sys=0.00, real=0.05 secs]
>> Total time for which application threads were stopped: 0.0480990 seconds
>> 2011-08-08T18:55:22.994+0000: 14142.407: [GC 14142.407: [ParNew:
>> 52844K->2014K(59008K), 0.0257020 secs] 1660793K->1609963K(4187776K)
>> icms_dc=0 , 0.0258530 secs] [Times: user=0.00 sys=0.00, real=0.03 secs]
>>
>>
>> I believe we are indeed using "virtualization"Š $0.68/hour, three nodes
>> 7 GB of memory
>> 20 EC2 Compute Units (8 virtual cores with 2.5 EC2 Compute Units each)
>> 1690 GB of instance storage
>> 64-bit platform
>> I/O Performance: High
>> API name: c1.xlarge
>>
>>
>> I am absolutely sure this is unpredictable and cause of a problem is
>>sharing
>> the same hardware with 3-4 other users. Box such as cc1.4xlarge is
>>shared
>> with 3 other users to make 3 instances of c1.xlarge, and no swap file
>>(but
>> "swap caching"?)
>> 23 GB of memory
>> 33.5 EC2 Compute Units (2 x Intel Xeon X5570, quad-core ³Nehalem²
>> architecture)
>> 1690 GB of instance storage
+
Jean-Daniel Cryans 2011-08-29, 18:58
+
Erik Onnen 2011-08-29, 18:52
+
Fuad Efendi 2011-08-29, 19:02