You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@jclouds.apache.org by Brock Noland <br...@cloudera.com> on 2013/07/21 20:42:38 UTC
Fwd: Question on how to configure retries/sleep time
Hi,
I am using jclouds with great success over at Apache Hive to implement our
testing infrastructure on top of ec2 spot instances. However, I have two
issues:
1) The polling service is simply too aggressive. If I have it on trace
logging I see it hitting the ec2 endpoint very quickly. I'd like to slow
the polling down by about 10x as my account is getting disabled due to
exceeding the request limit.
2) I'd like to increase both the sleep time and number of retries in
BackoffLimitedRetryHandler.
How do I make these configuration changes with JClouds? Do I create a
Guice module?
Brock
Re: Question on how to configure retries/sleep time
Posted by Everett Toews <ev...@RACKSPACE.COM>.
Hi Brock,
Those properties should be configurable if you're using the ComputeService. Have a look at [1] for an example.
HTH,
Everett
[1] https://github.com/jclouds/jclouds-examples/blob/master/rackspace/src/main/java/org/jclouds/examples/rackspace/cloudservers/CreateServer.java#L74-82
On Jul 21, 2013, at 2:22 PM, Brock Noland wrote:
Thanks, giving that a try. I also created https://issues.apache.org/jira/browse/JCLOUDS-197 as I think BackoffLimitedRetryHandler is a little too aggressive.
On Sun, Jul 21, 2013 at 1:52 PM, Andrew Bayer <an...@gmail.com>> wrote:
Rate limit on EC2 should improve in 1.6.2-incubating - I made a few changes that should cut down on the number of calls it makes, etc, that help from hitting the rate limit in the first place, and it's correctly using BackoffLimitedRetryHandler to delay/retry when it does hit rate limits. Give the 1.6.2-SNAPSHOT artifacts a try?
A.
On Sun, Jul 21, 2013 at 11:42 AM, Brock Noland <br...@cloudera.com>> wrote:
Hi,
I am using jclouds with great success over at Apache Hive to implement our testing infrastructure on top of ec2 spot instances. However, I have two issues:
1) The polling service is simply too aggressive. If I have it on trace logging I see it hitting the ec2 endpoint very quickly. I'd like to slow the polling down by about 10x as my account is getting disabled due to exceeding the request limit.
2) I'd like to increase both the sleep time and number of retries in BackoffLimitedRetryHandler.
How do I make these configuration changes with JClouds? Do I create a Guice module?
Brock
--
Apache MRUnit - Unit testing MapReduce - http://mrunit.apache.org<http://mrunit.apache.org/>
Re: Question on how to configure retries/sleep time
Posted by Brock Noland <br...@cloudera.com>.
Thanks, giving that a try. I also created
https://issues.apache.org/jira/browse/JCLOUDS-197 as I think
BackoffLimitedRetryHandler
is a little too aggressive.
On Sun, Jul 21, 2013 at 1:52 PM, Andrew Bayer <an...@gmail.com>wrote:
> Rate limit on EC2 should improve in 1.6.2-incubating - I made a few
> changes that should cut down on the number of calls it makes, etc, that
> help from hitting the rate limit in the first place, and it's correctly
> using BackoffLimitedRetryHandler to delay/retry when it does hit rate
> limits. Give the 1.6.2-SNAPSHOT artifacts a try?
>
> A.
>
>
> On Sun, Jul 21, 2013 at 11:42 AM, Brock Noland <br...@cloudera.com> wrote:
>
>>
>> Hi,
>>
>> I am using jclouds with great success over at Apache Hive to implement
>> our testing infrastructure on top of ec2 spot instances. However, I have
>> two issues:
>>
>> 1) The polling service is simply too aggressive. If I have it on trace
>> logging I see it hitting the ec2 endpoint very quickly. I'd like to slow
>> the polling down by about 10x as my account is getting disabled due to
>> exceeding the request limit.
>>
>> 2) I'd like to increase both the sleep time and number of retries in
>> BackoffLimitedRetryHandler.
>>
>> How do I make these configuration changes with JClouds? Do I create a
>> Guice module?
>>
>> Brock
>>
>>
>
--
Apache MRUnit - Unit testing MapReduce - http://mrunit.apache.org
Re: Question on how to configure retries/sleep time
Posted by Andrew Bayer <an...@gmail.com>.
Rate limit on EC2 should improve in 1.6.2-incubating - I made a few changes
that should cut down on the number of calls it makes, etc, that help from
hitting the rate limit in the first place, and it's correctly using
BackoffLimitedRetryHandler to delay/retry when it does hit rate limits.
Give the 1.6.2-SNAPSHOT artifacts a try?
A.
On Sun, Jul 21, 2013 at 11:42 AM, Brock Noland <br...@cloudera.com> wrote:
>
> Hi,
>
> I am using jclouds with great success over at Apache Hive to implement our
> testing infrastructure on top of ec2 spot instances. However, I have two
> issues:
>
> 1) The polling service is simply too aggressive. If I have it on trace
> logging I see it hitting the ec2 endpoint very quickly. I'd like to slow
> the polling down by about 10x as my account is getting disabled due to
> exceeding the request limit.
>
> 2) I'd like to increase both the sleep time and number of retries in
> BackoffLimitedRetryHandler.
>
> How do I make these configuration changes with JClouds? Do I create a
> Guice module?
>
> Brock
>
>