You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@trafficserver.apache.org by Adam Jefferiss <aj...@gmail.com> on 2011/06/24 11:06:51 UTC
Traffic Server 3.0 disk usage
Hi all,
I've been using web-polygraph[1] to run some performance tests through
traffic server 3.0 and have noticed that compared to 2.0.1 the disk usage is
considerably higher on average, for example one run has 2.0.1 marked down at
18% disk usage on average while 3.0 is up to 50%.
I've got logging set to errors only to test to see if the transaction
logging was causing the problem, but without much difference between the
two. I've also effectively disabled the cache in storage.conf by setting the
cache size to 0, again without too much of a difference.
Is there anything obvious that I've missed?
We're looking at having traffic server replace our current proxy, but the
disk usage results are causing a few concerns.
Any information you guys you can point me towards would be great.
Thanks,
Adam
[1] http://www.web-polygraph.org
Re: Traffic Server 3.0 disk usage
Posted by John Plevyak <jp...@gmail.com>.
I am not sure I understand this. The "cache" is supposed to run on raw disk
(preferred) or
on a preallocated file, so "disk usage" should be fixed.
john
On Mon, Jun 27, 2011 at 1:08 AM, Adam Jefferiss <aj...@gmail.com>wrote:
> Hi,
>
> Sorry about the late reply, the logging disk usage seems to be running on
> average higher. Although the cache disk usage does peak the highest.
>
>
> On 24 June 2011 17:48, Billy Vierra <bv...@sortatechie.com> wrote:
>
>> Adam,
>> Which disk usage is higher? Cache disk or logging disk?
>>
>>
>> On 06/24/2011 02:06 AM, Adam Jefferiss wrote:
>>
>>> Hi all,
>>>
>>> I've been using web-polygraph[1] to run some performance tests through
>>> traffic server 3.0 and have noticed that compared to 2.0.1 the disk usage is
>>> considerably higher on average, for example one run has 2.0.1 marked down at
>>> 18% disk usage on average while 3.0 is up to 50%.
>>>
>>> I've got logging set to errors only to test to see if the transaction
>>> logging was causing the problem, but without much difference between the
>>> two. I've also effectively disabled the cache in storage.conf by setting the
>>> cache size to 0, again without too much of a difference.
>>>
>>> Is there anything obvious that I've missed?
>>>
>>> We're looking at having traffic server replace our current proxy, but the
>>> disk usage results are causing a few concerns.
>>>
>>> Any information you guys you can point me towards would be great.
>>>
>>> Thanks,
>>> Adam
>>>
>>> [1] http://www.web-polygraph.org
>>>
>>
>
Re: Traffic Server 3.0 disk usage
Posted by Adam Jefferiss <aj...@gmail.com>.
Hi,
Sorry about the late reply, the logging disk usage seems to be running on
average higher. Although the cache disk usage does peak the highest.
On 24 June 2011 17:48, Billy Vierra <bv...@sortatechie.com> wrote:
> Adam,
> Which disk usage is higher? Cache disk or logging disk?
>
>
> On 06/24/2011 02:06 AM, Adam Jefferiss wrote:
>
>> Hi all,
>>
>> I've been using web-polygraph[1] to run some performance tests through
>> traffic server 3.0 and have noticed that compared to 2.0.1 the disk usage is
>> considerably higher on average, for example one run has 2.0.1 marked down at
>> 18% disk usage on average while 3.0 is up to 50%.
>>
>> I've got logging set to errors only to test to see if the transaction
>> logging was causing the problem, but without much difference between the
>> two. I've also effectively disabled the cache in storage.conf by setting the
>> cache size to 0, again without too much of a difference.
>>
>> Is there anything obvious that I've missed?
>>
>> We're looking at having traffic server replace our current proxy, but the
>> disk usage results are causing a few concerns.
>>
>> Any information you guys you can point me towards would be great.
>>
>> Thanks,
>> Adam
>>
>> [1] http://www.web-polygraph.org
>>
>
Re: Traffic Server 3.0 disk usage
Posted by Billy Vierra <bv...@sortatechie.com>.
Adam,
Which disk usage is higher? Cache disk or logging disk?
On 06/24/2011 02:06 AM, Adam Jefferiss wrote:
> Hi all,
>
> I've been using web-polygraph[1] to run some performance tests through
> traffic server 3.0 and have noticed that compared to 2.0.1 the disk
> usage is considerably higher on average, for example one run has 2.0.1
> marked down at 18% disk usage on average while 3.0 is up to 50%.
>
> I've got logging set to errors only to test to see if the transaction
> logging was causing the problem, but without much difference between
> the two. I've also effectively disabled the cache in storage.conf by
> setting the cache size to 0, again without too much of a difference.
>
> Is there anything obvious that I've missed?
>
> We're looking at having traffic server replace our current proxy, but
> the disk usage results are causing a few concerns.
>
> Any information you guys you can point me towards would be great.
>
> Thanks,
> Adam
>
> [1] http://www.web-polygraph.org