Thanks.,that helps...
So..what happens if I have in existing path on a smaller drive and I
attach a new volume and switch? In essence looking for best practice on
how ot handle this.
Is it better to do a rolling upgrade where I start new machines fresh with
1T then remove the smaller nodes 1 by
On 11/12/12 19:11, Kevin Burton wrote:
Thank you. I have restarted the server several times and I don't see riak
running with 'ps ax'. The configuration is set to come up on reboot so there
shouldn't be any other instance of riak around to run.
It's not necessarily another Riak instance but som
Thank you but this command didn't return anything:
[azureuser@bsicentos2 ~]$ sudo netstat -nap | grep LISTEN | grep -E
"8089|8099"
[sudo] password for azureuser:
[azureuser@bsicentos2 ~]$
The full "LISTENING" list is:
[azureuser@bsicentos2 ~]$ sudo netstat -nap | grep LISTEN
tcp0 0
Hi,
I'm new with Riak. I'm trying to use riak-erlang-client but I have some
problems.
I can connect with the browser to localhost:8091
http://localhost:8091/ping -> ok
But I can do a ping with the erlang client.
1> {ok, Pid} = riakc_pb_socket:start_link("127.0.0.1", 8091).
{ok,<0.34.0>}
2> riak
On 12/12/12 14:15, Kevin Burton wrote:
Thank you but this command didn't return anything:
OK, perhaps you're using the same port for different functions in Riak?
In particular, I see you're using 8099 for your http port but in my
setup 8099 is used for "handoff_port".
Here's the ports I'm u
Good catch! Thank you. If I have three machines that take up ports
8098,8099, and 8100 what do you suggest the handoff port should be? When is
it used?
-Original Message-
From: Shane McEwan [mailto:sh...@mcewan.id.au]
Sent: Wednesday, December 12, 2012 8:25 AM
To: Kevin Burton
Cc: riak-u
I am guessing from the comments that this port should be the same for all
machines in the cluster?
%% riak_handoff_port is the TCP port that Riak uses for
%% intra-cluster data handoff.
{handoff_port, 8099 },
-Original Message-
From: Shane McEwan
solved
I just had to connect to the Protobuf Port instead of http port.
2012/12/12 Pablo Vieytes
> Hi,
> I'm new with Riak. I'm trying to use riak-erlang-client but I have some
> problems.
>
> I can connect with the browser to localhost:8091
>
> http://localhost:8091/ping -> ok
>
> But I can
You're pinging the protocol buffer port 8091. HTTP is on another port, check
your app.config setting.
Eric
On Dec 12, 2012, at 6:23 AM, Pablo Vieytes wrote:
> Hi,
> I'm new with Riak. I'm trying to use riak-erlang-client but I have some
> problems.
>
> I can connect with the browser to local
On 12/12/12 14:32, Kevin Burton wrote:
Good catch! Thank you. If I have three machines that take up ports
8098,8099, and 8100 what do you suggest the handoff port should be? When is
it used?
No problem!
Each machine can use the same ports as the other machines. You just
can't use the same po
Hey Pablo,
Are you trying to use the Erlang protobuf client on the http port? You
can validate which port riak is listening to by checking for pb_port
variable in your app.config[1]. By default protobuf listens on port
8087.
To resolve this we will need your app.config along with the output of ne
I was just following the docs and they seem to set a different port or each
node but that was probably because the "Quick Start" was on one machine.
Thanks for the clarification.
Now I am running into a different problem. I joined one machine to the
cluster just fine. But joining the third node to
The join command needs to be given a node name and IP of a machine
already in the cluster.
So if you have machines:
riak@10.79.110.52
riak@10.79.108.25
riak@10.79.90.11
Then pick one to be the "master" (e.g. riak@10.79.110.52) and on the
other two machines run:
sudo riak-admin cluster join
My "master" is node1 which in vm.args as the node name of riak@10.79.108.25.
So from the "master" (riak@10.79.108.25) I issue the command 'sudo
riak-admin cluster join riak@10.79.108.15' (essentially joining to itself) I
get:
[azureuser@bsicentos1 ~]$ sudo riak-admin cluster join riak@10.79.108.25
There's no need to join the "master" to itself. Just join the other
nodes to the master.
On 12/12/12 15:33, Kevin Burton wrote:
My "master" is node1 which in vm.args as the node name of riak@10.79.108.25.
So from the "master" (riak@10.79.108.25) I issue the command 'sudo
riak-admin cluster join
So again from the master I do 'sudo riak-admin cluster join
riak@10.79.90.11' but I get:
[sudo] password for azureuser:
Attempting to restart script through sudo -H -u riak
Failed: This node is already a member of a cluster
Looking at the plan:
[azureuser@bsicentos1 ~]$ sudo riak-admi
Run the join commands on the OTHER machines . . . NOT the master.
Use the master's IP address in the join commands but the actual command
must be run on the non-master machines. You're telling the "slave"
machines to join the cluster by telling them which machine they need to
talk to to get th
So let me get it straight I have three machines:
riak@10.79.108.25 - The master
riak@10.79.110.52
riak@10.79.90.11
Since I have already joined riak@10.79.108.25 to riak@10.79.110.52. I only
have to go to riak@10.79.90.11 and do 'sudo riak-admin cluster join
riak@10.79.108.25'? Does it matter wher
So now I have
[azureuser@bsicentos3 ~]$ sudo riak-admin member-status
Attempting to restart script through sudo -H -u riak
= Membership
==
Status RingPendingNode
Looks good to me!
On 12/12/12 16:10, Kevin Burton wrote:
So now I have
[azureuser@bsicentos3 ~]$ sudo riak-admin member-status
Attempting to restart script through sudo -H -u riak
= Membership
==
Status RingPendingNode
On last question. Since this is a cluster there is only one public IP
address. Doesn't that necessitate the ports being different values (ie 8088
-> node1, 8089 -> node2, 8090 -> node3 or something like that)?
-Original Message-
From: Shane McEwan [mailto:sh...@mcewan.id.au]
Sent: Wednesd
You have to understand that there is no master node in Riak, all nodes are
equal, so you have as many entrypoint IPs as servers:
http://IP1:8089/riak
http://IP2:8089/riak
http://IP3:8089/riak
Etc...
You can query data, put data, delete data, etc... In any node at any time
Rohman
Sent from my i
But when setting up a cluster on VM's there is only one public address. So
in your scenario there is only IP not IP1, IP2, IP3. So they are all equal
including the IP address which addresses the cluster. The public ports
direct traffic to individual machines. At least that is the way I understand
i
Well, I don't work with VMs, for me 1 node = 1 full server
Sent from my iPhone
On 12/12/2012, at 17:43, "Kevin Burton" wrote:
> But when setting up a cluster on VM's there is only one public address. So
> in your scenario there is only IP not IP1, IP2, IP3. So they are all equal
> including the
Looking at
http://docs.basho.com/riak/latest/tutorials/installation/Installing-on-Windo
ws-Azure/ It looks like the setup is more of the later (one public IP, two
public ports) Although the documentation doesn't create a public port for
pb. So given this setup I am assuming that we only have this o
Hey everyone,
I'm currently using riak_kv as a reference of how to implement riak_core
and see that whenever a new coordinator process is needed, a new one is
created via their supervisor. But in the case of the get and put
coordinators, the supervisor is not used.
Besides eliminating the su
If your writes and reads are written with PR+PW > N (note PR/PW and not R/W),
then you should get read-your-writes consistency. There is unfortunately a bug
with PR/PW [1] that means this is not a 100% guarantee at the moment. I'm
hoping for it to be fixed by Riak 1.4, if not earlier. I don't ha
On 12 Dec 2012, at 19:20, David Fox wrote:
> Hey everyone,
>
> I'm currently using riak_kv as a reference of how to implement riak_core and
> see that whenever a new coordinator process is needed, a new one is created
> via their supervisor. But in the case of the get and put coordinators, th
Alright, so when creating your own coordinators, it's probably best to
use the supervisor until you see it as a bottleneck. Correct?
On 12/12/2012 13:36, Russell Brown wrote:
On 12 Dec 2012, at 19:20, David Fox wrote:
Hey everyone,
I'm currently using riak_kv as a reference of how to implem
OJ Reeves and I are happy to announce that CorrugatedIron (the Riak .NET
client) has reached 1.0. The full announcement is available at
http://corrugatediron.org/blog/v1/, but the highlights are fluent
interfaces for everyone, Riak v1.2.1 feature compatibility, better .NET
Task Parallel Library sup
Woot! Awesome work. Thanks for the hard work, gents.
Mark
On Wed, Dec 12, 2012 at 12:08 PM, Jeremiah Peschka
wrote:
> OJ Reeves and I are happy to announce that CorrugatedIron (the Riak .NET
> client) has reached 1.0. The full announcement is available at
> http://corrugatediron.org/blog/v1/, bu
Thanks Reid, we have that implementation for Read Your Writes already. We'd
like to implement Monotonic Read Consistency as well. For Monotonic Read
Consistency, we were hoping to do Vector Clock comparison but we don't know how
to convert the byte[] into different client's timestamps.
Here's a
On Dec 12, 2012, at 4:01 PM, "Vergara, Jeaneth Aguilar"
wrote:
> Thanks Reid, we have that implementation for Read Your Writes already. We'd
> like to implement Monotonic Read Consistency as well. For Monotonic Read
> Consistency, we were hoping to do Vector Clock comparison but we don't know
Kevin,
First off I want to make sure when you say "public" you don't actually mean
"public to the internet"? If so, we suggest keeping Riak behind a firewall
or VPN and have your application(s) talk to it through those access points.
Also, many people use a proxy like [HAProxy](http://haproxy.1w
Hey David,
My comments are inline.
On Wednesday, December 12, 2012 at 3:38 AM, David Montgomery wrote:
> Thanks.,that helps...
>
> So..what happens if I have in existing path on a smaller drive and I attach
> a new volume and switch?
- Stop riak.
- Copy the data out.
- Mount the new v
Monotonic read consistency means if a process has seen a particular value for
the object, any subsequent accesses will never return any previous values.
(quoted from
http://www.allthingsdistributed.com/2007/12/eventually_consistent.html).
Yes, the cache is local to each actor. The actor can rea
Hi all,
I'd just like to follow up Jeremiah's email with a couple of links.
Package on Nuget: http://nuget.org/packages/CorrugatedIron/
New docs on 2i usage:
http://corrugatediron.org/documentation/SecondaryIndexes.html
New docs on fluent search usage:
http://corrugatediron.org/documentation/Riak
Glad to do it. Lots of great work you've guys have done.
-Adron
On Wed, Dec 12, 2012 at 4:07 PM, OJ Reeves wrote:
> Hi all,
>
> I'd just like to follow up Jeremiah's email with a couple of links.
>
> Package on Nuget: http://nuget.org/packages/CorrugatedIron/
> New docs on 2i usage:
> http://c
38 matches
Mail list logo
