You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@karaf.apache.org by XiLai Dai <xl...@talend.com> on 2012/02/09 10:52:57 UTC
About cluster:node-ping command in Cellar
Hi,
When I try to run cluster:node-ping command with Karaf + Cellar cluster env, it always give a reply with about 10 seconds delay time if the target node is on the other machine.
karaf@trun> cluster:node-ping 192.168.2.104:5701
Pinging node 192.168.2.104:5701
PING 1 192.168.2.104:5701 10002ms
PING 2 192.168.2.104:5701 10001ms
PING 3 192.168.2.104:5701 10002ms
PING 4 192.168.2.104:5701 10000ms
Is it Ok or any configuration need to be set? Or it's a known issue. Thanks!
Regards.
Xilai
Re: About cluster:node-ping command in Cellar
Posted by Ioannis Canellos <io...@gmail.com>.
Hi Xilai,
The output is misleading, due to a bug. What actually happens is that all commands timeout after 10 seconds, but the ping command does not properly handle the timeout resulting in this false output.
I think that this issue happens when pinging nodes on separate groups. Could you please file a jira about it?
On 9 Φεβ 2012, at 12:25 μ.μ., Jean-Baptiste Onofré wrote:
> Hi Xilai,
>
> Some questions:
>
> - Which Cellar version ?
> - Which kind of discover used (static, unicast/multicast) ?
>
> If you do a system ping, what kind of latency do you have ?
>
> I tried on my VM without problem.
>
> Regards
> JB
>
> On 02/09/2012 10:52 AM, XiLai Dai wrote:
>> Hi,
>>
>> When I try to run cluster:node-ping command with Karaf + Cellar cluster env, it always give a reply with about 10 seconds delay time if the target node is on the other machine.
>>
>> karaf@trun> cluster:node-ping 192.168.2.104:5701
>> Pinging node 192.168.2.104:5701
>> PING 1 192.168.2.104:5701 10002ms
>> PING 2 192.168.2.104:5701 10001ms
>> PING 3 192.168.2.104:5701 10002ms
>> PING 4 192.168.2.104:5701 10000ms
>>
>> Is it Ok or any configuration need to be set? Or it's a known issue. Thanks!
>>
>> Regards.
>> Xilai
>
> --
> Jean-Baptiste Onofré
> jbonofre@apache.org
> http://blog.nanthrax.net
> Talend - http://www.talend.com
Ioannis Canellos
FuseSource
Blog: http://iocanel.blogspot.com
Apache Karaf Committer & PMC
Apache Camel Committer
Apache ServiceMix Committer
Apache Gora Committer
Apache DirectMemory Committer
Re: How to blacklist specific feature with cellar
Posted by Jean-Baptiste Onofré <jb...@nanthrax.net>.
http feature.
Let me explain deeper ;)
In fact, the [group].feature.[name].blacklist.[inbound|outbound]
property blocks the Cellar cluster event to be send (outbound) or
received (inbound) on an event handler. If you configure it after the
http feature installation, it's "too late", the events have been already
sent with features:install command (due to the FeaturesListener) (only
the new node will not be sync). So to block the http feature, you should
block the event sending before installing the http feature.
Regards
JB
On 02/14/2012 07:50 AM, XiLai Dai wrote:
> Thanks JB!
>
> "before the features install ..."
> Which feature did you mean? Cellar feature or http feature(those want to be blacklisted)?
>
> should we provide a predefined org.apache.karaf.cellar.groups.cfg file which include all features,configs want to be blacklisted before start karaf container with cellar?
>
> Thanks.
> Xilai
> -----Original Message-----
> From: Jean-Baptiste Onofré [mailto:jb@nanthrax.net]
> Sent: Tuesday, February 14, 2012 2:32 PM
> To: user@karaf.apache.org
> Subject: Re: How to blacklist specific feature with cellar
>
> Hi Xilai,
>
> The configuration of blacklist/whitelist should be done before the
> features install (else, it's assigned to the local node cluster group).
>
> Anyway, your configuration is correct.
>
> Regards
> JB
>
> On 02/14/2012 04:58 AM, XiLai Dai wrote:
>> Hi,
>>
>> Just want to make some specific feature (eg. http) disabled sync between the nodes.
>>
>> For example, start two karaf 2.2.5 containers
>>
>> Create etc/org.ops4j.pax.web.cfg for both of them.
>> Set port for first karaf: org.osgi.service.http.port=8080
>> Set port for second karaf: org.osgi.service.http.port=8081
>>
>> install http feature on both of them
>> karaf@root>features:install http
>>
>> then , install cellar on both of them
>> karaf@root>features:addurl mvn:org.apache.karaf.cellar/apache-karaf-cellar/2.2.3/xml/features
>> karaf@root>features:install cellar
>>
>> and add the http feature into the blacklist in the etc/org.apache.karaf.cellar.groups.cfg
>> default.features.blacklist.inbound=cellar,http
>> default.features.blacklist.outbound=cellar,http
>>
>> then , restart 2 containers.
>>
>> What expected will be both http://localhost:8080 and http://localhost:8081 could be listening, but unfortunately, only http://localhost:8081 can be accessible.
>> What else configuration/steps have to do? Or it's a blacklist problem of cellar?
>>
>> Thanks.
>> Xilai
>
--
Jean-Baptiste Onofré
jbonofre@apache.org
http://blog.nanthrax.net
Talend - http://www.talend.com
RE: How to blacklist specific feature with cellar
Posted by XiLai Dai <xl...@talend.com>.
Thanks JB!
"before the features install ... "
Which feature did you mean? Cellar feature or http feature(those want to be blacklisted)?
should we provide a predefined org.apache.karaf.cellar.groups.cfg file which include all features,configs want to be blacklisted before start karaf container with cellar?
Thanks.
Xilai
-----Original Message-----
From: Jean-Baptiste Onofré [mailto:jb@nanthrax.net]
Sent: Tuesday, February 14, 2012 2:32 PM
To: user@karaf.apache.org
Subject: Re: How to blacklist specific feature with cellar
Hi Xilai,
The configuration of blacklist/whitelist should be done before the
features install (else, it's assigned to the local node cluster group).
Anyway, your configuration is correct.
Regards
JB
On 02/14/2012 04:58 AM, XiLai Dai wrote:
> Hi,
>
> Just want to make some specific feature (eg. http) disabled sync between the nodes.
>
> For example, start two karaf 2.2.5 containers
>
> Create etc/org.ops4j.pax.web.cfg for both of them.
> Set port for first karaf: org.osgi.service.http.port=8080
> Set port for second karaf: org.osgi.service.http.port=8081
>
> install http feature on both of them
> karaf@root>features:install http
>
> then , install cellar on both of them
> karaf@root>features:addurl mvn:org.apache.karaf.cellar/apache-karaf-cellar/2.2.3/xml/features
> karaf@root>features:install cellar
>
> and add the http feature into the blacklist in the etc/org.apache.karaf.cellar.groups.cfg
> default.features.blacklist.inbound=cellar,http
> default.features.blacklist.outbound=cellar,http
>
> then , restart 2 containers.
>
> What expected will be both http://localhost:8080 and http://localhost:8081 could be listening, but unfortunately, only http://localhost:8081 can be accessible.
> What else configuration/steps have to do? Or it's a blacklist problem of cellar?
>
> Thanks.
> Xilai
--
Jean-Baptiste Onofré
jbonofre@apache.org
http://blog.nanthrax.net
Talend - http://www.talend.com
Re: How to blacklist specific feature with cellar
Posted by Jean-Baptiste Onofré <jb...@nanthrax.net>.
Hi Xilai,
The configuration of blacklist/whitelist should be done before the
features install (else, it's assigned to the local node cluster group).
Anyway, your configuration is correct.
Regards
JB
On 02/14/2012 04:58 AM, XiLai Dai wrote:
> Hi,
>
> Just want to make some specific feature (eg. http) disabled sync between the nodes.
>
> For example, start two karaf 2.2.5 containers
>
> Create etc/org.ops4j.pax.web.cfg for both of them.
> Set port for first karaf: org.osgi.service.http.port=8080
> Set port for second karaf: org.osgi.service.http.port=8081
>
> install http feature on both of them
> karaf@root>features:install http
>
> then , install cellar on both of them
> karaf@root>features:addurl mvn:org.apache.karaf.cellar/apache-karaf-cellar/2.2.3/xml/features
> karaf@root>features:install cellar
>
> and add the http feature into the blacklist in the etc/org.apache.karaf.cellar.groups.cfg
> default.features.blacklist.inbound=cellar,http
> default.features.blacklist.outbound=cellar,http
>
> then , restart 2 containers.
>
> What expected will be both http://localhost:8080 and http://localhost:8081 could be listening, but unfortunately, only http://localhost:8081 can be accessible.
> What else configuration/steps have to do? Or it's a blacklist problem of cellar?
>
> Thanks.
> Xilai
--
Jean-Baptiste Onofré
jbonofre@apache.org
http://blog.nanthrax.net
Talend - http://www.talend.com
How to blacklist specific feature with cellar
Posted by XiLai Dai <xl...@talend.com>.
Hi,
Just want to make some specific feature (eg. http) disabled sync between the nodes.
For example, start two karaf 2.2.5 containers
Create etc/org.ops4j.pax.web.cfg for both of them.
Set port for first karaf: org.osgi.service.http.port=8080
Set port for second karaf: org.osgi.service.http.port=8081
install http feature on both of them
karaf@root>features:install http
then , install cellar on both of them
karaf@root>features:addurl mvn:org.apache.karaf.cellar/apache-karaf-cellar/2.2.3/xml/features
karaf@root>features:install cellar
and add the http feature into the blacklist in the etc/org.apache.karaf.cellar.groups.cfg
default.features.blacklist.inbound=cellar,http
default.features.blacklist.outbound=cellar,http
then , restart 2 containers.
What expected will be both http://localhost:8080 and http://localhost:8081 could be listening, but unfortunately, only http://localhost:8081 can be accessible.
What else configuration/steps have to do? Or it's a blacklist problem of cellar?
Thanks.
Xilai
Re: About cluster:node-ping command in Cellar
Posted by Jean-Baptiste Onofré <jb...@nanthrax.net>.
Hi Xilai,
after checking in the code, in 2.2.3, the timeout handling is not fully
fixed.
So I think that you raise the timeout.
I advice you:
- to check if you don't have Windows firewall/firewall blocking the
multicast/hazelcast messages
- to switch in static host definition in the
etc/org.apache.karaf.cellar.instance.cfg and etc/hazelcast.xml
I raised:
https://issues.apache.org/jira/browse/KARAF-1194
I will fix this issue for Cellar 2.2.4.
Regards
JB
On 02/10/2012 03:14 AM, XiLai Dai wrote:
> Hi, JB,
>
> The Cellar version is 2.2.3.
> features:addurl mvn:org.apache.karaf.cellar/apache-karaf-cellar/2.2.3/xml/features
> features:install cellar
>
> I've made nothing change to the cellar configuration files in the etc/, keep the default value in the etc/ org.apache.karaf.cellar.instance.cfg :
> multicastEnabled=true
> multicastGroup=224.2.2.3
> multicastPort=54327
> multicastTimeoutSeconds=2
>
> so the discovery type should be multicast.
>
> I've tested with a system ping (Win 7), the replies is quickly at most about hundreds milliseconds. But the replies from Cellar ping are constantly about 10 seconds.
>
> Thanks
> Xilai
>
> -----Original Message-----
> From: Jean-Baptiste Onofré [mailto:jb@nanthrax.net]
> Sent: Thursday, February 09, 2012 6:25 PM
> To: user@karaf.apache.org
> Subject: Re: About cluster:node-ping command in Cellar
>
> Hi Xilai,
>
> Some questions:
>
> - Which Cellar version ?
> - Which kind of discover used (static, unicast/multicast) ?
>
> If you do a system ping, what kind of latency do you have ?
>
> I tried on my VM without problem.
>
> Regards
> JB
>
> On 02/09/2012 10:52 AM, XiLai Dai wrote:
>> Hi,
>>
>> When I try to run cluster:node-ping command with Karaf + Cellar cluster env, it always give a reply with about 10 seconds delay time if the target node is on the other machine.
>>
>> karaf@trun> cluster:node-ping 192.168.2.104:5701
>> Pinging node 192.168.2.104:5701
>> PING 1 192.168.2.104:5701 10002ms
>> PING 2 192.168.2.104:5701 10001ms
>> PING 3 192.168.2.104:5701 10002ms
>> PING 4 192.168.2.104:5701 10000ms
>>
>> Is it Ok or any configuration need to be set? Or it's a known issue. Thanks!
>>
>> Regards.
>> Xilai
>
--
Jean-Baptiste Onofré
jbonofre@apache.org
http://blog.nanthrax.net
Talend - http://www.talend.com
RE: About cluster:node-ping command in Cellar
Posted by XiLai Dai <xl...@talend.com>.
Hi, JB,
The Cellar version is 2.2.3.
features:addurl mvn:org.apache.karaf.cellar/apache-karaf-cellar/2.2.3/xml/features
features:install cellar
I've made nothing change to the cellar configuration files in the etc/, keep the default value in the etc/ org.apache.karaf.cellar.instance.cfg :
multicastEnabled=true
multicastGroup=224.2.2.3
multicastPort=54327
multicastTimeoutSeconds=2
so the discovery type should be multicast.
I've tested with a system ping (Win 7), the replies is quickly at most about hundreds milliseconds. But the replies from Cellar ping are constantly about 10 seconds.
Thanks
Xilai
-----Original Message-----
From: Jean-Baptiste Onofré [mailto:jb@nanthrax.net]
Sent: Thursday, February 09, 2012 6:25 PM
To: user@karaf.apache.org
Subject: Re: About cluster:node-ping command in Cellar
Hi Xilai,
Some questions:
- Which Cellar version ?
- Which kind of discover used (static, unicast/multicast) ?
If you do a system ping, what kind of latency do you have ?
I tried on my VM without problem.
Regards
JB
On 02/09/2012 10:52 AM, XiLai Dai wrote:
> Hi,
>
> When I try to run cluster:node-ping command with Karaf + Cellar cluster env, it always give a reply with about 10 seconds delay time if the target node is on the other machine.
>
> karaf@trun> cluster:node-ping 192.168.2.104:5701
> Pinging node 192.168.2.104:5701
> PING 1 192.168.2.104:5701 10002ms
> PING 2 192.168.2.104:5701 10001ms
> PING 3 192.168.2.104:5701 10002ms
> PING 4 192.168.2.104:5701 10000ms
>
> Is it Ok or any configuration need to be set? Or it's a known issue. Thanks!
>
> Regards.
> Xilai
--
Jean-Baptiste Onofré
jbonofre@apache.org
http://blog.nanthrax.net
Talend - http://www.talend.com
Re: About cluster:node-ping command in Cellar
Posted by Jean-Baptiste Onofré <jb...@nanthrax.net>.
Hi Xilai,
Some questions:
- Which Cellar version ?
- Which kind of discover used (static, unicast/multicast) ?
If you do a system ping, what kind of latency do you have ?
I tried on my VM without problem.
Regards
JB
On 02/09/2012 10:52 AM, XiLai Dai wrote:
> Hi,
>
> When I try to run cluster:node-ping command with Karaf + Cellar cluster env, it always give a reply with about 10 seconds delay time if the target node is on the other machine.
>
> karaf@trun> cluster:node-ping 192.168.2.104:5701
> Pinging node 192.168.2.104:5701
> PING 1 192.168.2.104:5701 10002ms
> PING 2 192.168.2.104:5701 10001ms
> PING 3 192.168.2.104:5701 10002ms
> PING 4 192.168.2.104:5701 10000ms
>
> Is it Ok or any configuration need to be set? Or it's a known issue. Thanks!
>
> Regards.
> Xilai
--
Jean-Baptiste Onofré
jbonofre@apache.org
http://blog.nanthrax.net
Talend - http://www.talend.com