You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@karaf.apache.org by sditlinger <st...@ditlinger.com> on 2015/02/07 00:01:29 UTC
KARAF-1315 still in Cellar 2.3.4?
I am seeing the behavior of bug KARAF-1315 in ServiceMix 5.4/Karaf
2.4.1/Cellar 2.3.4, yet this bug was reported fixed in Cellar 2.2.4. Should
it not be fixed in cellar 2.3.4 also?
My screen shot would exactly duplicate the description in KARAF-1315:
multiple Karaf instances spread across 2 different VMs
cluster:node-list ......shows all instances
cluster:group-create master .......succeeds
cluster:group-join master remote.machine.name:5701 ........times out
--
View this message in context: http://karaf.922171.n3.nabble.com/KARAF-1315-still-in-Cellar-2-3-4-tp4038365.html
Sent from the Karaf - User mailing list archive at Nabble.com.
Re: KARAF-1315 still in Cellar 2.3.4?
Posted by sditlinger <st...@ditlinger.com>.
I agree, it would seem to be a network issue.
Wasn't thinking it was a multicast problem, since cluster:node-list always
works from inside VM (i.e. lists all nodes, local and remote). Seems like
node-list would not work if multicast itself was the problem. Need to
figure out the difference between node-list and group-list in network comms.
thanks,
Steve
--
View this message in context: http://karaf.922171.n3.nabble.com/KARAF-1315-still-in-Cellar-2-3-4-tp4038365p4038466.html
Sent from the Karaf - User mailing list archive at Nabble.com.
Re: KARAF-1315 still in Cellar 2.3.4?
Posted by Achim Nierbeck <bc...@googlemail.com>.
Hi,
sounds to me more like a VM/Network Issue.
Per default Cellar/Hazelcast tries to communicate via Multicast.
So from the picture you describe you're having a network issue.
regards, Achim
2015-02-10 23:48 GMT+01:00 sditlinger <st...@ditlinger.com>:
> OK, continuing with the scientific method, I have now added a node running
> on
> a virtual machine to the mix. So that is two servicemix installations,
> each
> on a physical machine, and one servicemix installation on a virtual machine
> that is hosted by one of those physical machines.
>
> On the phyisical machines, I create a group; I can add remote members,
> including the node on the VM
>
> If I do a group-list on the physical machines, I see all members of all
> groups, including the member from the VM
>
> If I do a group-list on VM, it lists all groups, but all I see is the node
> local to the VM listed as a member
>
> hmmmm....
>
> VM issue or Cellar issue ?
>
>
>
> --
> View this message in context:
> http://karaf.922171.n3.nabble.com/KARAF-1315-still-in-Cellar-2-3-4-tp4038365p4038445.html
> Sent from the Karaf - User mailing list archive at Nabble.com.
>
--
Apache Member
Apache Karaf <http://karaf.apache.org/> Committer & PMC
OPS4J Pax Web <http://wiki.ops4j.org/display/paxweb/Pax+Web/> Committer &
Project Lead
blog <http://notizblog.nierbeck.de/>
Co-Author of Apache Karaf Cookbook <http://bit.ly/1ps9rkS>
Software Architect / Project Manager / Scrum Master
Re: KARAF-1315 still in Cellar 2.3.4?
Posted by sditlinger <st...@ditlinger.com>.
Mystery solved! Cause of problem found...
For reasons that one can intuit, if not immediately explain, Karaf/Cellar
installations on VMs are not able to resolve machine names to IP addresses
in the way physical machines are able to. The environment I am working in is
a lab network with no DNS. The Karaf/Cellar nodes on physical machines had
no problems forming and listing groups, node pinging, etc The VMs listed
only local nodes in groups, couldn't ping nodes outside VM.
So we added entries in the c:\Windows\System32\drivers\etc\hosts file for
all of our machines (physical and virtual) on the virtual machines - problem
solved!
--
View this message in context: http://karaf.922171.n3.nabble.com/KARAF-1315-still-in-Cellar-2-3-4-tp4038365p4038537.html
Sent from the Karaf - User mailing list archive at Nabble.com.
Re: KARAF-1315 still in Cellar 2.3.4?
Posted by sditlinger <st...@ditlinger.com>.
OK, here is where I am at after another full day of experimenting...
I have two physical machines, one of which hosts a VM
I have two instances of ServiceMix 5.4/Karaf 2.4.1/Cellar 2.3.4, on physical
machine 75 (PM75) and the other on virtual machine 80 (VM80)
On PM75, I can create groups and add members, including remote members
On VM80, adding remote members to groups yields a timeout, but the member is
added to group
On PM75, cluster:group-list shows all members of all groups, including
remote members
On VM80, cluster:group-list shows only local members of groups
On PM75, I can cluster:node-ping VM80 with success
On VM80, when I cluster:node-ping PM75, I get 10 timeouts
I ran wireshark on PM75 and VM80...
When I node-ping VM80 from PM75, I see the 10 ping messages and 10 responses
in wireshark on both machines
When I node-ping PM75 from VM80, I see the 10 ping messages on BOTH
machines, but see responses on NEITHER machine
What would cause cellar on PM75 to NOT respond to a ping message from VM80?
Is this relevant?: successful ping messages are 463 bytes in size and yield
an 805 byte response, the unsuccessful ping messages are 464 bytes in size
and of course yield no response
--
View this message in context: http://karaf.922171.n3.nabble.com/KARAF-1315-still-in-Cellar-2-3-4-tp4038365p4038509.html
Sent from the Karaf - User mailing list archive at Nabble.com.
Re: KARAF-1315 still in Cellar 2.3.4?
Posted by sditlinger <st...@ditlinger.com>.
Using Bridge for network on the VM. Gonna try some different VM configs
today.
Haven't tried tcp-ip yet. Wasn't thinking it was a multicast problem, since
cluster:node-list always works from inside VM (i.e. lists all nodes, local
and remote). If it can discover other nodes thru multicast for node-list,
why not for tracking group membership for group-list?
--
View this message in context: http://karaf.922171.n3.nabble.com/KARAF-1315-still-in-Cellar-2-3-4-tp4038365p4038464.html
Sent from the Karaf - User mailing list archive at Nabble.com.
Re: KARAF-1315 still in Cellar 2.3.4?
Posted by Jean-Baptiste Onofré <jb...@nanthrax.net>.
IMHO, it looks like a network issue. Which kind of network do you use on
the VM: bridge or NAT ?
Did you try to use static tcp-ip instead of multicast ?
Regards
JB
On 02/10/2015 11:48 PM, sditlinger wrote:
> OK, continuing with the scientific method, I have now added a node running on
> a virtual machine to the mix. So that is two servicemix installations, each
> on a physical machine, and one servicemix installation on a virtual machine
> that is hosted by one of those physical machines.
>
> On the phyisical machines, I create a group; I can add remote members,
> including the node on the VM
>
> If I do a group-list on the physical machines, I see all members of all
> groups, including the member from the VM
>
> If I do a group-list on VM, it lists all groups, but all I see is the node
> local to the VM listed as a member
>
> hmmmm....
>
> VM issue or Cellar issue ?
>
>
>
> --
> View this message in context: http://karaf.922171.n3.nabble.com/KARAF-1315-still-in-Cellar-2-3-4-tp4038365p4038445.html
> Sent from the Karaf - User mailing list archive at Nabble.com.
>
--
Jean-Baptiste Onofré
jbonofre@apache.org
http://blog.nanthrax.net
Talend - http://www.talend.com
Re: KARAF-1315 still in Cellar 2.3.4?
Posted by sditlinger <st...@ditlinger.com>.
OK, continuing with the scientific method, I have now added a node running on
a virtual machine to the mix. So that is two servicemix installations, each
on a physical machine, and one servicemix installation on a virtual machine
that is hosted by one of those physical machines.
On the phyisical machines, I create a group; I can add remote members,
including the node on the VM
If I do a group-list on the physical machines, I see all members of all
groups, including the member from the VM
If I do a group-list on VM, it lists all groups, but all I see is the node
local to the VM listed as a member
hmmmm....
VM issue or Cellar issue ?
--
View this message in context: http://karaf.922171.n3.nabble.com/KARAF-1315-still-in-Cellar-2-3-4-tp4038365p4038445.html
Sent from the Karaf - User mailing list archive at Nabble.com.
Re: KARAF-1315 still in Cellar 2.3.4?
Posted by sditlinger <st...@ditlinger.com>.
Yes, local VM, and yes it is VirtualBox.
--
View this message in context: http://karaf.922171.n3.nabble.com/KARAF-1315-still-in-Cellar-2-3-4-tp4038365p4038465.html
Sent from the Karaf - User mailing list archive at Nabble.com.
Re: KARAF-1315 still in Cellar 2.3.4?
Posted by Jean-Baptiste Onofré <jb...@nanthrax.net>.
Interesting. By VM, you mean "local" VM (using VirtualBox for instance) ?
Regards
JB
On 02/10/2015 08:37 PM, sditlinger wrote:
> OK, if I attempt the same action (i.e., creating a group and adding remote
> node to that group) with two different physical machines (no VMs), it works
> (remote machine is added to group). There is something peculiar about VMs
> that is causing problem.
>
> Steve
>
>
>
> --
> View this message in context: http://karaf.922171.n3.nabble.com/KARAF-1315-still-in-Cellar-2-3-4-tp4038365p4038443.html
> Sent from the Karaf - User mailing list archive at Nabble.com.
>
--
Jean-Baptiste Onofré
jbonofre@apache.org
http://blog.nanthrax.net
Talend - http://www.talend.com
Re: KARAF-1315 still in Cellar 2.3.4?
Posted by sditlinger <st...@ditlinger.com>.
OK, if I attempt the same action (i.e., creating a group and adding remote
node to that group) with two different physical machines (no VMs), it works
(remote machine is added to group). There is something peculiar about VMs
that is causing problem.
Steve
--
View this message in context: http://karaf.922171.n3.nabble.com/KARAF-1315-still-in-Cellar-2-3-4-tp4038365p4038443.html
Sent from the Karaf - User mailing list archive at Nabble.com.
Re: KARAF-1315 still in Cellar 2.3.4?
Posted by sditlinger <st...@ditlinger.com>.
cluster:group-set master remote.machine.name:5701 .....also times out
Interesting that I created group on one VM and it showed up in the list of
groups in Karaf instances on both VMs with no problem. I can add local
Karaf node to that group from both VMs, but cluster:group-list only shows
local member of group from either VM
--
View this message in context: http://karaf.922171.n3.nabble.com/KARAF-1315-still-in-Cellar-2-3-4-tp4038365p4038366.html
Sent from the Karaf - User mailing list archive at Nabble.com.
Re: KARAF-1315 still in Cellar 2.3.4?
Posted by Jean-Baptiste Onofré <jb...@nanthrax.net>.
HI,
Thanks for the report, I didn't reproduce in Karaf vanilla.
Let me take a look in SMX.
Regards
JB
On 02/07/2015 12:01 AM, sditlinger wrote:
>
> I am seeing the behavior of bug KARAF-1315 in ServiceMix 5.4/Karaf
> 2.4.1/Cellar 2.3.4, yet this bug was reported fixed in Cellar 2.2.4. Should
> it not be fixed in cellar 2.3.4 also?
>
> My screen shot would exactly duplicate the description in KARAF-1315:
> multiple Karaf instances spread across 2 different VMs
>
> cluster:node-list ......shows all instances
>
> cluster:group-create master .......succeeds
>
> cluster:group-join master remote.machine.name:5701 ........times out
>
>
>
> --
> View this message in context: http://karaf.922171.n3.nabble.com/KARAF-1315-still-in-Cellar-2-3-4-tp4038365.html
> Sent from the Karaf - User mailing list archive at Nabble.com.
>
--
Jean-Baptiste Onofré
jbonofre@apache.org
http://blog.nanthrax.net
Talend - http://www.talend.com