Hi Peter,

A: Because it confuses the reader.
Q: Why?
A: No.
Q: Should I write my response above the quoted reply?

..so please use proper quoting as I'm doing below!

On Jun 20 2016, Peter Auyeung <[email protected]> wrote:
>> > s3qlstat on google storage: 
>> > Preparing test data... 
>> > Measuring throughput to cache... 
>> > Cache throughput with   4 KiB blocks: 12198 KiB/sec 
>> > Cache throughput with   8 KiB blocks: 24942 KiB/sec 
>> > Cache throughput with  16 KiB blocks: 49939 KiB/sec 
>> > Cache throughput with  32 KiB blocks: 85977 KiB/sec 
>> > Cache throughput with  64 KiB blocks: 132026 KiB/sec 
>> > Cache throughput with 128 KiB blocks: 187159 KiB/sec 
>> > Measuring raw backend throughput.. 
>> > Backend throughput: 41466 KiB/sec 
>> > Test file size: 0.00 MiB 
>>
>> Maybe try with a bigger test file? 
>>
>> If the data you are storing is distributed over similarly small files, 
>> then it is possible that the bad performance is due to latency (as 
>> opposed to bandwidth). For every file, S3QL needs to make at least one 
>> PUT request. So if you have millions of tiny files, it will spend a lot 
>> of time just waiting for the server to reply without ever being able to 
>> fully utilize the bandwidth. 
>
> Yes it is exactly the case.....
>
> Wonder if there is a better way to tackle this

At some indefinite point in the future, S3QL will be able to store
multiple files in one storage object
(https://bitbucket.org/nikratio/s3ql/issues/5/support-fragments). Until
then, the best workaround is to work with bigger files (e.g. tar them
first).

Best,
-Nikolaus

-- 
GPG encrypted emails preferred. Key id: 0xD113FCAC3C4E599F
Fingerprint: ED31 791B 2C5C 1613 AF38 8B8A D113 FCAC 3C4E 599F

             »Time flies like an arrow, fruit flies like a Banana.«

-- 
You received this message because you are subscribed to the Google Groups 
"s3ql" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to