Best unofficial Apache Server developers community |
Home >> discussions >> Project Voldemort discussions
Project Voldemort discussions Recent threads
| Threads | Replies | First post | Last post | |
|---|---|---|---|---|
|
Automatic Balancing of Cluster By: Srikanth Is "Automatic Balancing of cluster" on addition of new nodes part of
the road-map? May be use Zookeeper similar to "kafka"
(http://incubator.apache.org/ kafka/design.html) to enable automatic
cluster management? Regards, Srikanth
|
0 | May 24 2012 14:22 |
May 24 2012 14:22 |
|
|
ANN] YobiDrive By: Francois We are launching soon (this is in private beta for 3 months)
YobiDrive: If you need collaborative PKI encryption, smart access
right management, fully elastic storage with NO SQL db at all, S3 and
folder structure support, and optional multiple…
|
0 | May 24 2012 04:51 |
May 24 2012 04:51 |
|
|
Ann] virtual clouddrive By: Maarten Koopmans Http://www.virtual-clouddrive.org Data emancipation for the rest of
us. Feel free to ask for help - this is important to me on a moral
level. Based on Voldemort for the metadata stores that actually
protect you. Enjoy, Maarten
|
5 | May 23 2012 13:09 |
May 24 2012 04:36 |
|
|
Trying to reconfigure my Voldemort cluster to use zones. By: ctasada Hi everyone, Right now I've a Voldemort cluster configured and
running. Due to different changes we need to reconfigure it to work
in 2 datacenters. My requirement are: I need the information
replicated between zones and between nodes in the…
|
21 | May 16 2012 11:00 |
May 21 2012 09:45 |
|
|
Client thread safety, memory leaks? By: Maarten Koopmans Hi, The list is a bit quiet - I hope somebody can shine light on
this - even if I'm totally off track here. I have a multi-threaded
server that acts as a client to Voldemort, The threads are more like
actors/processes, in that they share nothing…
|
7 | Apr 4 2012 09:58 |
May 14 2012 13:51 |
|
|
New store / Build in concurrency resolver By: Francois After looking at cache store (a nice implementation, that we like
because hash makes sense), we are building a new type of store called
DiskMap , for 1k to 5k/s Ops/s in write and > 10 k Ops/s in read
on cheap machines with a very low ratio…
|
7 | Mar 17 2012 12:03 |
May 13 2012 02:15 |
|
|
IPv6 support By: Maarten Koopmans Hi, Quick but (more and more) important question: does Voldemort
support IPv6 out of the box? I assume it does, currently have no
access to an IPv6 box but will have to deploy to one in the near
future. --Maarten
|
2 | Mar 14 2012 10:48 |
May 10 2012 21:23 |
|
|
Distributed Cluster and Firewall By: Srikanth Hi, We have a 4 node cluster, with 2 nodes in one DC and two nodes
in another DC. Nodes in each DC have been assigned a zone id, and
hence, we have two zones configured. I have setup the 4 nodes in full
replication. We are observing that after…
|
1 | May 9 2012 13:30 |
May 10 2012 00:01 |
|
|
Routing type differences By: ctasada Hi, I'm trying to understand the difference between the client and
server routing (as in stores.xml <store>/<routing>) I see
that the options are client or server and I found this description in
one of the wiki pages Routing can be…
|
6 | Apr 30 2012 07:55 |
May 1 2012 14:29 |
|
|
How to configure zones on cluster.xml? By: BJ2 I'm trying to configure 3 zones on cluster.xml. I've read your sample
like following. <zone>
<zone-id>0</zone-id>
<proximity-list>1</proximity-list> <zone>
<zone> …
|
3 | Mar 6 2012 20:55 |
May 1 2012 00:29 |
|
|
when does voldemort perform its repairs on nodes for data? By: S Ahmed When a client inserts a key/value, I believe it is up to the caller
to determine how many servers it wants to write to. If you have 3
servers, and you only write to a single server, voldemort will have to
eventually write to the rest of the…
|
3 | Apr 5 2012 22:26 |
Apr 29 2012 19:52 |
|
|
Very OT] Detecting socket closes By: Maarten Koopmans Hi, This post is completely off-topic, but the root of a problem I'm
running into lies close to that of a sophisticated NoSQL store written
in Java (it uses Voldemort as NoSQL metadata store). I am wrapping up
(and soon will be releasing on…
|
0 | Apr 16 2012 13:32 |
Apr 16 2012 13:32 |
|
|
Lost data By: Francois Hello, We are facing a strange case on a test cluster. 4 nodes in 2
zones, a few thousands of partitions (simulate a young cluster ready
to scale with machines of different sizes). Read/W quorums: 2/2,
replication factor 3 Node 4 stopped, in…
|
1 | Apr 4 2012 11:01 |
Apr 4 2012 11:04 |
|
|
Project issues status By: ctasada Hi everyone, As I have already mentioned in some other emails on
this group, I'm recently in charge of a project that's using
Voldemort as a backend. We're facing some issues, mainly performance,
that I'm trying to solve. Anyway, while looking…
|
3 | Mar 8 2012 16:47 |
Apr 4 2012 10:07 |
|
|
Best Configuration for a Production Enviorment By: ctasada Hi everyone, Since some weeks ago I'm involved in a project that's
using project-voldemort as a backend. We're having some performance
issues due to an increment in the system load. For what I've been
reading in the group and checking the code…
|
8 | Mar 1 2012 12:04 |
Mar 20 2012 11:06 |
|
|
Replication doesn't seem to work in 0.90.1 By: Steve I have a test case with 2 nodes using BDB, replication factor 2,
required writes 2. I a key/value into a node 1 StoreClient, then pull
down node 1. I read the key from a node 2 StoreClient, and assert that
it's there, but it's not. The same test…
|
11 | Feb 1 2012 01:10 |
Mar 6 2012 20:55 |
|
|
Voldemort with the new BDB JE 5.0.34 By: erdody Hello, We are thinking about migrating to BDB JE 5.0.34, based on
this patch:
http://code.google.com/p/project-voldemort/issues/detail?id=373 We
experienced extremely low rebalancing performance and we are hoping
DiskOrderedCursor will make it…
|
2 | Mar 6 2012 18:11 |
Mar 6 2012 19:32 |
|
|
Presentation By: Francois Hello, Is this the right place to organize a quick presentation of
the ongoing or planned Voldemort-based projects ? Just for everyone to
know who's doing what with it, and what could be the contributions to
the core. Francois
|
1 | Feb 11 2012 04:26 |
Feb 22 2012 01:40 |
|
|
Hinted Handoff question By: liulei For each failed put operation, use a specified HintedHandoffStrategy
to route the request to a node elsewhere in the cluster and perform a
synchronous put to the slop store on that node. For data consistency,
I think client should read data from…
|
3 | Feb 20 2012 04:36 |
Feb 22 2012 01:04 |
|
|
how to config voldemort not to bound on 0.0.0.0 By: wonderyl i would like voldemort bound to a specific ip, like 10.1.1.1 not
0.0.0.0 or * could anyone teach me how? thanks
|
2 | Feb 1 2012 01:10 |
Feb 7 2012 14:30 |