incubator-cassandra-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From aaron morton <aa...@thelastpickle.com>
Subject Re: Need to run nodetool repair on a cluster running 1.1.6 if no deletes
Date Sun, 23 Dec 2012 19:33:35 GMT
> I'm seeing some conflicting advice out there about whether you need to run nodetool repair
within GCGraceSeconds with 1.x. Can someone clarify two things:
Can you point me to the advice not to run it. 

> (1) Do I need to run repair if I'm running 1.x?
Yes. 
Repair / Anti Entropy needs to be run so that deletes are distributed. 

> (2) Should I bother running repair if I don't have any deletes? Anything drawbacks to
not running it?
IMHO Yes. 
It's is the way to ensure our data is consistency distributed. Hinted Handoff and Read Repair
are optimisations which reduce the probability of getting a Digest Mismatch (inconsistent
data) when reading at a CL > ONE. 

I also find it makes it easier to do maintenance and deal with potential emergencies if you
have confidence the data is full distributed. 

However, other than distributed deletes, I do not know of any other technical issues for running
repair. And I have seen a lot of systems running without out. (I just prefer a very boring
life). 

Cheers 

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

@aaronmorton
http://www.thelastpickle.com

On 23/12/2012, at 5:33 AM, Andrew Bialecki <andrew.bialecki@gmail.com> wrote:

> Hey everyone,
> 
> I'm seeing some conflicting advice out there about whether you need to run nodetool repair
within GCGraceSeconds with 1.x. Can someone clarify two things:
> 
> (1) Do I need to run repair if I'm running 1.x?
> (2) Should I bother running repair if I don't have any deletes? Anything drawbacks to
not running it?
> 
> Thanks,
> Andrew
> 


Mime
View raw message