Can you mail the list the output of the attached test Makefile? If the
builtin gmake CURDIR variable isn't being set, there isn't much we can do
for you, I'm afraid. It means that your gmake build is woefully broken or
something, somewhere is managing to set it to an empty string.
--Ryan
On Wed
I've not messed with partition numbers, so I'm going to take a stab and say 64
across 3 hosts.
Andrew
On 13/05/2010, at 1:34 PM, Sean Cribbs wrote:
The chunk size-splitting makes sense, although it is irritating to hack around
it. How many partitions are you using in your cluster?
Sean Crib
The chunk size-splitting makes sense, although it is irritating to hack around
it. How many partitions are you using in your cluster?
Sean Cribbs
Developer Advocate
Basho Technologies, Inc.
http://basho.com/
On May 12, 2010, at 11:26 PM, Andrew Harvey wrote:
> I believe it's a curb bug.
>
>
I believe it's a curb bug.
curl --raw gives me expected output, but the chunks yielded by the
@client.http.get are not the same chunks shown by curl – they're fragmented as
shown earlier in this thread., that said the chunks returned are pretty huge
(1.2 MiB), which might explain why curb is cu
I haven't tested net/http, I couldn't get it to work when passing a block to
keys. I am pretty sure this is the same issue I mentioned on 5/4 and you
reported as a bug. It only seems to happen with large numbers of keys. Riak
will return partial JSON objects and they are rejected in ripple be
If the problem is with net/http then we need to fix that backend, not apply
hacks to the generic portions. You can verify that Riak will return valid JSON
objects in each chunk by adding --raw to the curl command. Here's what I get
after loading the Fast Track stock data: http://gist.github.co
This is a confirmed bug and will be fixed soon. You can check out
http://github.com/adamhunter/ripple/blob/associations/lib/riak/bucket.rb : 65
for a temporary hack.
Thanks,
Adam
On May 12, 2010, at 10:17 PM, Andrew Harvey wrote:
> I've just run into this one as well.
>
>
> Not sure if it'
I've just run into this one as well.
Not sure if it's a curb thing or what, but it's making me a sad panda. We have
500k+ keys in our bucket, and I'm being returned chunks of invalid json so
nothing is being yielded to my block. Needless to say, the task I'm doing here
I'd prefer to do in a Map
Well I don't know if that's normal or not. I just know from erlang experience
that DETS is super-slow, especially when restoring or backup up etc.
On May 12, 2010, at 6:04 PM, Germain Maurice wrote:
> I'm using Innostore (10-1) and riak_0.10.1 on an ubuntu 9.10.
>
>
> Le 13/05/2010 00:28, Alan
I'm using Innostore (10-1) and riak_0.10.1 on an ubuntu 9.10.
Le 13/05/2010 00:28, Alan McConnell a écrit :
I get similar behavior using the innostore storage engine. In
addition, beam.smp memory footprint balloons to using all available
system memory.
I'm running latest binary installs of
I get similar behavior using the innostore storage engine. In addition,
beam.smp memory footprint balloons to using all available system memory.
I'm running latest binary installs of riak and innostore (0.10.1) on ubuntu
8.04.
On Wed, May 12, 2010 at 3:07 PM, Preston Marshall wrote:
> What stor
(resent: I failed at using gmail for lists)
Wow, fast replies! Thank you Alexander and Grant.
Your instructions worked and I was able to start a 2nd node on my
workstation here (on a different cluster). Last time I tried this I was
using our dev server so I still have some investigating to do, bu
In order to run multiple nodes on a host, and consequently multiple clusters,
you will need to edit the vm.args and app.config for each embedded node (an
embedded node being a copy of the rel/riak directory you built). For each of
these nodes you will need to change the "-name r...@127.0.0.1" s
What storage engine are you using?
On May 12, 2010, at 4:57 PM, Germain Maurice wrote:
> Hi,
>
> I'm testing the backup process of a node and all nodes.
> I'm very surprised by the time it took.
>
> The beginning using 'date' :
> # date ; ls -lh
> mardi 11 mai 2010, 14:31:54 (UTC+0200)
> total 8
Check the ports in app.config and in the vm config in the etc riak/
etc. Check the tutorial in fast track riak. But I think the main thing
that determines chit chat within a cluster is the channel name which
is "riak", ie. r...@127.0.0.1, by default.
@siculars on twitter
http://siculars.pos
Hi,
I'm testing the backup process of a node and all nodes.
I'm very surprised by the time it took.
The beginning using 'date' :
# date ; ls -lh
mardi 11 mai 2010, 14:31:54 (UTC+0200)
total 8,0K
-rw-r--r-- 1 riak riak8 2010-05-11 14:31 backup-20100511
drwxr-xr-x 2 riak riak 4,1K 2010-05-05 1
Hi all,
Trying to configure 2 clusters on the same machine. The 2nd node I fire up
quits after a second or two. Is this possible, and I'm just missing some
configuration? Or should I be spinning up VMs for additional clusters?
tia,
Sami
___
riak-users m
Hi,
Celine is my name. i am 25yrs old and i live in Dakar Senegal.
it is very great when i get your email id and i felt a great spark in my heart despite that i have not seen you in person but in my heart it realy gives me a nice pleasure to communicate with you. You can reach me via email fo
For some reason the standard GNU make CURDIR isn't being set. Are you sure
you're using a GNU make? What does gmake --version return?
--Ryan
On Tue, May 11, 2010 at 9:58 PM, kg9020 wrote:
> Hello,
>
> Update here is the error
> Running make -C c_src
> tar -xzf nsprpub-4.8.tar.gz
> (cd /nsprpu
Sorry, I'm having lots of difficulty even getting a Free/Dragonfly/NetBSD setup
that will let me attempt to build riak. I'm afraid I can't help you at this
time.
Sean Cribbs
Developer Advocate
Basho Technologies, Inc.
http://basho.com/
On May 12, 2010, at 1:40 AM, kg9020 wrote:
> Hello every
Ya, I had the same problem. You also need to include the vclock. It's
tricky. Notice without the client id the vclock grows. Basically when
you don't use a client id riak randomly assigns one.
I think I settled on creating a new key without a client id and then
using a client id on updates
Hi all,
>From the wiki:
Client ID
All requests should include the X-Riak-ClientId header, which can be
any string that uniquely identifies the client, for purposes of
tracing object modifications in the vector clock.
A few days ago I added it to riak-js (http://github.com/frank06/riak-js):
opti
Afternoon, Evening, Morning -
A quick recap for today: a few bits from IRC and an announcement about
an email address.
Best,
Mark
Community Manager
wiki.basho.com
twitter.com/pharkmillups
-
Riak Recap for 5/11
1) Q --- Is there anyway to update the riak metadata fields without
rewriting
23 matches
Mail list logo
