Another point, I've done another test on my 5 node cluster.
Created a keyspace with replication factor of 5 and inserted some data in
it.
Run a full repair on each node to make sstable appear on disk.
Then run multiple times on each nodes:
1 - nodetool repair
2 - nodetool repair -pr
Due to
Thanks for your reply,
But can't figure out why it's not recommendend, by DataStax, to run
primary-range with incremental repair...
It's just doing less work on each repair call on the repaired node.
At the end, when all the nodes are repaired using either method, all data
is equally consistent?
The objective of non-incremental primary-range repair is to avoid redoing
work, but with incremental repair anticompaction will segregate repaired
data so no extra work is done on the next repair.
You should run nodetool repair [ks] [table] in all nodes sequentially. The
more often you run, the
On 2016-10-24 13:39 (-0500), Alexander Dejanovski
wrote:
> Hi Sean,
>
> In order to mitigate its impact, anticompaction is not fully executed when
> incremental repair is run with -pr. What you'll observe is that running
> repair on all nodes with -pr will leave
Thanks.
Sean
From: Alexander Dejanovski [a...@thelastpickle.com]
Sent: Monday, October 24, 2016 10:39 AM
To: user@cassandra.apache.org
Subject: Re: incremental repairs with -pr flag?
Hi Sean,
In order to mitigate its impact, anticompaction is not fully executed
Hi Sean,
In order to mitigate its impact, anticompaction is not fully executed when
incremental repair is run with -pr. What you'll observe is that running
repair on all nodes with -pr will leave sstables marked as unrepaired on
all of them.
Then, if you think about it you realize it's no big
Hey,
In the datastax documentation on repair [1], it says,
"The partitioner range option is recommended for routine maintenance. Do
not use it to repair a downed node. Do not use with incremental repair
(default for Cassandra 3.0 and later)."
Why is it not recommended to use -pr with