I tend to caution against making very large batch mutations or multi gets, by 
which I mean 100's of rows at a time. 

Each row request becomes a task and they can temporarily fill the mutation or 
read thread pool. Meaning overall *client* request throughout drops while a big 
request is chewed through.  

This this is more of an issue with smaller clusters. As Dean says, the client 
request is performed in parallel on multiple machines. 

Cheers

-----------------
Aaron Morton
Freelance Cassandra Developer
New Zealand

@aaronmorton
http://www.thelastpickle.com

On 12/12/2012, at 3:03 AM, "Hiller, Dean" <[email protected]> wrote:

> Each node is doing it's thing in parallel….they on purpose do NOT co-ordinate 
> as they do not need to so each one is doing it's scan on the rows it has 
> individually.
> 
> If all rows "happen" to be on the same server, sure some may be done 
> sequentially depending on number of rows vs. thread pool size.
> 
> As far as a single row is concerned, I know mutations to a single row are 
> serialised as Aaron has said as much but you are talking about multiple rows 
> here.
> 
> Later,
> Dean
> 
> From: Wei Zhu <[email protected]<mailto:[email protected]>>
> Reply-To: "[email protected]<mailto:[email protected]>" 
> <[email protected]<mailto:[email protected]>>, Wei Zhu 
> <[email protected]<mailto:[email protected]>>
> Date: Monday, December 10, 2012 3:15 PM
> To: Cassandr usergroup 
> <[email protected]<mailto:[email protected]>>
> Subject: Re: multiget_slice SlicePredicate
> 
> Well, not sure how parallel is multiget. Someone is saying it's in parallel 
> sending requests to the different nodes and on each node it's executed 
> sequentially. I didn't bother looking into the source code yet. Anyone knows 
> it for sure?
> 
> I am using Hector, just copied the thrift definition from Cassandra site for 
> reference.
> 
> You are right, the count is for each individual row.
> 
> Thanks.
> -Wei
> 
> ________________________________
> From: "Hiller, Dean" <[email protected]<mailto:[email protected]>>
> To: "[email protected]<mailto:[email protected]>" 
> <[email protected]<mailto:[email protected]>>; Wei Zhu 
> <[email protected]<mailto:[email protected]>>
> Sent: Monday, December 10, 2012 1:13 PM
> Subject: Re: multiget_slice SlicePredicate
> 
> What's wrong with multiget…parallel performance is great from multiple disks 
> and so usually that is a good thing.
> 
> Also, something looks wrong, since you have list<binary> keys, I would expect 
> the Map to be Map<binary, list<ColumnOrSuperColumn>>
> 
> Are you sure you have that correct?  IF you set range to 100, it should be 
> 100 columns each row but it never hurts to run the code and verify.
> 
> Later,
> Dean
> PlayOrm Developer
> 
> 
> From: Wei Zhu 
> <[email protected]<mailto:[email protected]><mailto:[email protected]<mailto:[email protected]>>>
> Reply-To: 
> "[email protected]<mailto:[email protected]><mailto:[email protected]<mailto:[email protected]>>"
>  
> <[email protected]<mailto:[email protected]><mailto:[email protected]<mailto:[email protected]>>>,
>  Wei Zhu 
> <[email protected]<mailto:[email protected]><mailto:[email protected]<mailto:[email protected]>>>
> Date: Monday, December 10, 2012 2:07 PM
> To: Cassandr usergroup 
> <[email protected]<mailto:[email protected]><mailto:[email protected]<mailto:[email protected]>>>
> Subject: multiget_slice SlicePredicate
> 
> I know it's probably not a good idea to use multiget, but for my use case, 
> it's the only choice,
> 
> I have question regarding the SlicePredicate argument of the multiget_slice
> 
> 
> The SlicePredicate takes slice_range which takes start, end and range. I 
> suppose start and end will apply to each individual row. How about range, is 
> it a accumulative column count of all the rows or to the individual row?
> If I set range to 100, is it 100 columns per row, or total?
> 
> Thanks for you reply,
> -Wei
> 
> multiget_slice
> 
> *
> map<string,list<ColumnOrSuperColumn>> multiget_slice(list<binary> keys, 
> ColumnParent column_parent, SlicePredicate predicate, ConsistencyLevel 
> consistency_level)
> 
> 
> 
> 

Reply via email to