Is this for testing purposes?
I think there is a flaw in your design. As Dustin said, your machine is not
going to be happy about 5000 concurrent IO operations unless you have 5000
cores.
Java's concurrent package is very easy to use and has low overhead-- take a
look at ExecutorService. Maybe
Hi, theese threads simulate users. I have jboss AS that manages all user
requests and for every request i need to update memcache. I will try spy
client tomorow.
On Thu, Feb 25, 2010 at 7:48 PM, Dustin wrote:
>
> On Feb 25, 7:21 am, mariyan nenchev wrote:
> > What will happen if i create 5000 t
Comment #4 on issue 123 by h...@terminal-consulting.de: Cross compile
OpenWrt
http://code.google.com/p/memcached/issues/detail?id=123
I'll try to setup a buildbot with openwrt trunk. By the way using the sdk
needs less
work.
--
You received this message because you are listed in the owne
Indeed, I was somehow wondering "why a Q for no reply?" ;-)
Anyway, it makes sense. I guess my issue is fully client side then.
cheers,
Jean-Charles
On 25 fév, 18:38, Dustin wrote:
> On Feb 25, 6:29 am, JC wrote:
>
> > trying to investigate an issue on the replicated delete of
> > libmemcached,
Comment #3 on issue 123 by dsallings: Cross compile OpenWrt
http://code.google.com/p/memcached/issues/detail?id=123
That's a great start. I'd love a slave if you can contribute one:
http://code.google.com/p/memcached/wiki/BuildFarm
That way, we can make sure we continue to support your config
Comment #2 on issue 123 by h...@terminal-consulting.de: Cross compile
OpenWrt
http://code.google.com/p/memcached/issues/detail?id=123
Yes. I'll setup a machine you can login an try to rebuild yourself with the
OpenWrt-SDK-x86.
This includes the toolchain for the release i try to build.
Is tha
Comment #1 on issue 123 by dsallings: Cross compile OpenWrt
http://code.google.com/p/memcached/issues/detail?id=123
Can you supply flags, toolchains, etc... Perhaps a build slave?
--
You received this message because you are listed in the owner
or CC fields of this issue, or because you starre
On Feb 25, 7:21 am, mariyan nenchev wrote:
> What will happen if i create 5000 threads(5000 users update simultaniosly)
> to update the cache just once at the same time?
Unless your box has 5,000 cores, your application is going to fall
apart in all kinds of other exciting ways here. It's cer
Status: New
Owner:
Labels: Type-Defect Priority-Medium
New issue 123 by h...@terminal-consulting.de: Cross compile OpenWrt
http://code.google.com/p/memcached/issues/detail?id=123
What steps will reproduce the problem?
1. Try to issue configure over the OpenWrt Build System
2. Cross Compile
your problem might indeed be related to the danga client.. if you take
a look at its connection pool (SockIOPool i believe) it has a lot of
synchronized code, which isn't very happy in a highly concurrent
environment. i recommend at least doing a test with the spy client to
see if it actually help
I assume you are running memcached in any kind of unix.
So do you had the java virtual machine and the memcached runing in the
box while you got those errors?. If so, forget about the network. But
you could check how busy is your box when you start those threads
(mem, cpu)
On the other hand, see
Every thing is one one machine in development, but it will be separate
server for the memcache with LAN.
Yes i tried it. What should i see? It stop updating the cache while i get
the exceptions above.
On Thu, Feb 25, 2010 at 6:31 PM, Carlos Alvarez wrote:
> That's a problem, because a lot of issu
That's a problem, because a lot of issues have to do with clients.
Anyway, how healthy is your network?.
Have you tried running memcached with -vv (two 'v' and not a 'w' and)
and see the log in the server?
On Thu, Feb 25, 2010 at 1:56 PM, mariyan nenchev
wrote:
> Hi, it is not up to me to chang
Hi, it is not up to me to change the client.
First thing I would suggest: use a modern client. The Danga client is
getting old, but the Spy client is up-to-date.
http://code.google.com/p/spymemcached/
- Nelz
On Thu, Feb 25, 2010 at 07:21, mariyan nenchev
wrote:
> What will happen if i create 5000 threads(5000 users update simultaniosly)
>
What will happen if i create 5000 threads(5000 users update simultaniosly)
to update the cache just once at the same time?
OK,
i wrote simple test to show you what is happening. I create 100 threads, and
i start them at the same time. They begin to update the cache.
public class UpdateableObject implements Serializable {
public int i = 0;
}
static int i = 0;
public static void main(String[] args) {
Hi,
Before 4 months i had to create cache for data stream messages coming from
TCP stream about 400 per second every second. I had to keep the last message
only. I implemented MDB that received the messages and updated the memcache
but it couldn't take the pressure. May be i didn't configured it c
We have a few thousands updates per second per box with memcached
(linux with eight core each) and the box hardly reports any activity.
The only problem we had was the terrible slow .net serialization and
the md5 computed when we plug hibernate with memcached.
In my experience, the bottlenecks at
Hi,
trying to investigate an issue on the replicated delete of
libmemcached, I found out that the no reply delete command is replying
something in case of error (as "key not found"), which makes it pretty
hard for the client library to follow the messages count and not work
as expected in certain
On Feb 25, 12:39 am, mariyan nenchev
wrote:
> Well this is not actually true. I tried to update memcache before with about
> 10 messages per second every second and it began to throw some IO
> exceptions. I am using theese classes:
> import com.danga.MemCached.MemCachedClient;
> import com.danga.
Well this is not actually true. I tried to update memcache before with about
10 messages per second every second and it began to throw some IO
exceptions. I am using theese classes:
import com.danga.MemCached.MemCachedClient;
import com.danga.MemCached.SockIOPool;
The other code that manages my cac
22 matches
Mail list logo