I rebuilt cluster again and still "node_puts_total": 0 on another node.
Found explanation for "node_puts": "Number of puts coordinated by this node in
the last minute"
However, the actual "puts" were sent to another node, not to this one, so I
assume this value does not represent puts between nodes, just direct puts to the
node.
I think the proper value to check would be "vnode_puts".
Box A = "vnode_puts": 2144
Box B = "vnode_puts": 2170
so it does seem it's OK
vlatko
-------- Original Message --------
Subject: Re: Very, very slow data insert on two node cluster
From: Eric Moritz <[email protected]>
To: Vlatko Basic <[email protected]>
Cc: [email protected]
Date: 06.05.2011 20:22
The fact that the other node is idle clues makes me think that the
node you're talking to is not reaching the other node. Both nodes
should be participating in the storage of the data.
On Fri, May 6, 2011 at 5:38 AM, Vlatko Basic<[email protected]> wrote:
Hi,
I've build a cluster with two physical Ubuntu 10.10 64-bit 4GB machines
(laptop, 2 cores, 2.4GHz and desktop, 4 cores, 3.1GHz) and I am trying to
"load_data goog.csv" (around 1500 records) from a Fast Track tutorial. It
took over 15 minutes.
Riak installation (both from DEB package) is correct and working. Other
examples worked as shown.
1. Inserts are very, very slow. Maybe 1 insert in 2 secs??? Memory used
around 50%. Network is wired.
2. All inserts go to the local node that I'm working on, the other node is
idle.
I attached status for both nodes and bucket properties.
Is it a bug, or have I made a bug? :-)
Thank for any ideas...
vlatko
_______________________________________________
riak-users mailing list
[email protected]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
_______________________________________________
riak-users mailing list
[email protected]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
