can not telnet membase 11210 port after I restart the service
I started use Membase serval days ago,every thing is perfect~
But, after restarting my service today,I found membase can not work.
telnet 192.168.1.199 11211
>stats
>SERVER_ERROR proxy write to downstream
telnet 192.168.1.199 11210
>port 11210:connected failed
netstat -an
TCP 192.168.1.199:139 0.0.0.0:0 LISTENING
TCP 192.168.1.199:445 192.168.1.100:6539 ESTABLISHED
TCP 192.168.1.199:445 192.168.1.102:52628 ESTABLISHED
TCP 192.168.1.199:1951 81.169.142.213:5938 ESTABLISHED
TCP 192.168.1.199:3389 192.168.1.103:3074 ESTABLISHED
TCP 192.168.1.199:11211 192.168.1.199:3620 TIME_WAIT
notice that:11210 port is not in this list
i have checked windows services,the MembaseServer service is running.and in taskmanager ,task "memcached" is running too.
It seems that Moxi work well and Membase does not work.
need help
I have shut down the win2003's firewall,and there is not any other running security software.
I have found a way to solve this problem.
Every time I start the service, I should run the following .bat file:
service_register.bat
service_stop.bat
service_start.bat
Then,everything is ok~
I am still trying to find what causes this problem.
Hi Simon, sorry for the delay in getting back to you. The problem is likely related to your IP address changing after the node has been setup. Do you have DHCP running in the environment?
Take a look at this post and let me know if it helps: http://techzone.couchbase.com/forums/thread/issues-dhcp-development-laptop
Perry
Thanks Perry~
My server running Membase service has a fixed Ip address, and DHCP running in the route not in my server.
And unfortunately, today ,I found even though running the three '.bat' files the moxi still couldnot connect to the Membase server,11210 port is not in the result of running 'netstat -an' in cmd.
simomo
Still me~~
I found a strange thing.
After running Service_stop.bat,I do find :
UDP 0.0.0.0:11210 *:*
in the result of 'netstat -an'.But 11211 and 8091 is not in this result.
And,after running Service_start.bat,I find :
TCP 0.0.0.0:8091 0.0.0.0:0 LISTENING
TCP 0.0.0.0:11211 0.0.0.0:0 LISTENING
in the result of 'netstat -an'.But 11210 is not in this result.
How can this happen?
It seems Moxi and Membase is mutually exclusive.
Simon, with the service stopped, can you figure out what is running on 11210?
'netstat -anb' should show the process that is listening on that port...
Perry
service stopped:
TCP 0.0.0.0:11210 0.0.0.0:0 LISTENING 2988
[memcached.exe]
service start:
TCP 0.0.0.0:8091 0.0.0.0:0 LISTENING 2388
[erl.exe]
TCP 0.0.0.0:11211 0.0.0.0:0 LISTENING 3968
[moxi.exe]
TCP 0.0.0.0:21100 0.0.0.0:0 LISTENING 2388
[erl.exe]
TCP 127.0.0.1:4599 0.0.0.0:0 LISTENING 3968
[moxi.exe]
TCP 127.0.0.1:4627 127.0.0.1:4626 ESTABLISHED 3968
[moxi.exe]
TCP 127.0.0.1:4628 127.0.0.1:4629 ESTABLISHED 3968
[moxi.exe]
TCP 127.0.0.1:4629 127.0.0.1:4628 ESTABLISHED 3968
[moxi.exe]
TCP 127.0.0.1:4630 127.0.0.1:4631 ESTABLISHED 3968
[moxi.exe]
TCP 127.0.0.1:4631 127.0.0.1:4630 ESTABLISHED 3968
[moxi.exe]
TCP 127.0.0.1:4633 127.0.0.1:8091 ESTABLISHED 3968
[moxi.exe]
TCP 127.0.0.1:8091 127.0.0.1:4633 ESTABLISHED 2388
[erl.exe]
So many maxi port~
The memcached process should not be running while the service is stopped. There's either a bug with our package, or something else is running memcached outside of the Membase service. It almost looks like there are two installations of Membase that are both trying to run...can you look for that as well?
Also, in your previous post I see that something is listening on the UDP port of 11210 whereas this last post shows it on the TCP port.
Can you manually kill all those processes and try starting the service up again?
Hi,perry
Thanks so much for your help.
My Membase server is an development Server,which shutdown everyday.
This server'env is windows2003server.I have not setup an single memcached in it before I install the Membase.
Because of the Membase problem,I have installed and uninstalled the Membase for several times,is it possible that uninstall operation does not clear up the files of the Membase, and this result of files run a memcached process?
The uninstall process "should" clear everything up, but it's possible that it's not.
After you uninstall, do you still see anything running on ports 11210, 11211, 4369 or 8091?
Perry
After I stopped the MembaseService ,the process 'memcached' is still running.
After I uninstalled the Membase, the process 'memcached' is not running.
This show that the uninstall process can clear everything ,but the process 'memcached' is not under MembaseService's control.
I reinstalled the membase again,now it works well~
I suspect there might be something else installed there since in my lab, the MembaseService stops the memcached process as well.
Glad you got it working, let me know if you run into any problems in the future.
Perry
The process memcached is out of control again.
After stop MembaseService:
process memcached is running.
UDP 0.0.0.0:11210 *:* 1552
[memcached.exe]
in 'netstat -anb' result
After start MembaseService:
process memcached is still running
None about memcached in 'netstat -anb' result
We round a circle and are ath the beginniing agian~
Perry, May it fix this that I use the Enterprise Version instead of Community Version?
simomo
I can't yet say what version will fix this without knowing what the issue is.
However, I can certainly say that if you purchase an Enterprise License, we will be able to work with you directly to ensure that this is resolved quickly.
To continue the investigation here, can you start the service and grab the output of http://:8091/diag
Look through there for any errors coming from the memcached process.
Perry
Hello,
Did you managed to the problem you encountered with the memcached proccess?
i have the same problem like you. after running the tests mentioned above i saw that the memcached service is not listening on port 11210.
In the diag report I can see the following error message about the memcached:
ERROR REPORT <0.268.0> 2011-04-12 14:04:34
===============================================================================
ns_1@***.***.***.***:ns_memcached:374: Unable to connect: {error,
{badmatch,
{error,
econnrefused}}}, retrying.
I saw that if I kill the process through "Process Explorer" and restart it not under the erl service then memcached start listening and all is working.
p.s
my server is running on windows server 2003 with service pack 2
Thanks in advance,
Roi
Hi,roiah
I use an pure memcache bucket instead of membase bucket.
In most of case,Membase service works well,but,yesterday,Membase service was out of control again.
I shutdown the Membase service,and terminal the memcached process in task-manager,restart the Membase service,then
everything is ok~
I guess,just guess, membase service has some bugs on controlling memcached process.
Hope this is helpful for you.
simomo2010
I use an pure memcache bucket instead of membase bucket.
In most of case,Membase service works well,but,yesterday,Membase service was out of control again.
I shutdown the Membase service,and terminal the memcached process in task-manager,restart the Membase service,then
everything is ok~
I guess,just guess, membase service has some bugs on controlling memcached process.
Hope this is helpful for you.
simomo2010
Thanks for your reply,
Unfortunately, your solution does not work for me.
I did managed to see in process explorer that when I stop the Membase service, memcached process is starting to work correctly.
I try to investigate it further more. will notify if I'll manage to fix the problem.
Regards,
Roi
Roi, are you saying that you stopped the Membase service and the memcached process starts working again?
It really sounds like there is some other memcached process running here...is that incorrect?
Perry
Would you post a solution here
I have been seeing the same issue
memcached.exe stops listening after i restart MembaseServer.
if started it from the command line:
memcached.exe -p 11210 -v -E "C:\P
rogram Files\Membase\Server\bin\bucket_engine\bucket_engine.so"
I would get:
Initialized SASL.
Loaded engine: Bucket engine v0.2
Supplying the following features: Multi tenancy
144: sasl result code: -1
144: Unknown sasl response: -1
Rafal
You'll actually need a bit more on the command-line to start this memcached properly:
memcached.exe –vvv –p –E /ep.so –r –e “dbname=/default-data/default”
I have your logs and am working with our engineers to diagnose. As I mentioned, we don't officially support Windows 2003 and so we need to spend a bit more time reproducing that environment. It would be really helpful to know whether this is working properly for you on a more recent operating system.
Perry
Hi Perry,
This is working perfectly on 2008 Win Server
Had some issues on 2003 Win Server Virtual Machine but finally got it working
Still not working on all 3 machines - 2003 Win Server (Physical Box). We do have 3 network interfaces on each of these boxes , not sure if that is making any difference. To make sure there is no ip conflict or so I disabled 2 adapters on one of these boxes but again no luck
Thanks,
Rafal
Rafal, can you describe what you had to do with the 2003 VM that you finally got working?
I encountered a similar issue with Membase 1.7.2 on Windows Server 2003. After a server restart, the machine would not respond on memcache buckets ports. Restarting the membase service would resolve the issue until the next machine restart.
On a hunch, I rigged a "delayed start" for the membase service, using the method described here: http://support.microsoft.com/kb/193888 . Windows Server 2003 does not have the "delayed start" option that services do on 2008. In my particular case, I added the W3SVC as a "DependOnService" key in order for it to delay loading membase until the web server was completely loaded.
We've had a few server restarts since this change was made and so far membase has come up correctly each time. I just wanted to share in case this helps anyone else!
env: win2003 32bit