Sorry I don't think this contains much of an architecture and implementation view.
It does not for example state that: - dtmd is a service started very early on each node in the cluster - dtmd supervision - the dtmd service uses multicast/broadcast(?) to discover peers - how does message send work - how does message broadcast work - once peers have been discovered, the service database is synced (?) - a part of the MDS library talks to the dtmd service using a UNIX socket - etc. Btw TIPC does no longer have any 3pp dependencies since it is part of the kernel. But in some distributions the kernel module might not be available. Thanks, Hans On 2015-04-09 06:32, [email protected] wrote: > osaf/services/infrastructure/dtms/README | 171 > +++++++++++++++++++++++++++++++ > 1 files changed, 171 insertions(+), 0 deletions(-) > > > diff --git a/osaf/services/infrastructure/dtms/README > b/osaf/services/infrastructure/dtms/README > new file mode 100644 > --- /dev/null > +++ b/osaf/services/infrastructure/dtms/README > @@ -0,0 +1,171 @@ > +# > +# -*- OpenSAF -*- > +# > +# (C) Copyright 2010 The OpenSAF Foundation > +# > +# This program is distributed in the hope that it will be useful, but > +# WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY > +# or FITNESS FOR A PARTICULAR PURPOSE. This file and program are licensed > +# under the GNU Lesser General Public License Version 2.1, February 1999. > +# The complete license can be accessed from the following location: > +# http://opensource.org/licenses/lgpl-license.php > +# See the Copying file included with the OpenSAF distribution for full > +# licensing terms. > +# > +# Author(s): Oracle > +# > + > +GENERAL > + > +This directory (osaf/services/infrastructure/dtms) contains an > +implementation of the infrastructure DTM service whic provides > +alternate TCP transport that is going to be used for the MDS > +discovery and Transport Mechanism. > + > +The advantages of choosing this Alternate transport for MDS > +Discovery mechanism based on the TCP/IP is that it doesn`t > +required any third-party modules dependency and the TCP work > + across the sub-nets (discovers multiple sub-net nodes). > + > +DESIGN > +Following are the are provided by MDS Discovery and Transport module (MDTM): > + > +Node Discovery: This is the way by which MDTM will report that > +an NODE (Cluster Member) has joined/left the cluster. MDTM will > +report the UP/DOWN of the NODE, once the Applications have subscribed > + to the NODE Subscription through the MDS. > + > +Service Discovery: This is the way by which MDTM will report the UP/DOWN > +of the services installed with MDS. MDS will report the UP/DOWN of the > +services, once the Applications have subscribed to the destination > +service through the MDS. > + > +Message Send: This is the way by which MDS delivers the messages sent by > +the applications to the destined service in the cluster using the MDTM. > + > +Message Receive: This is the way by which MDS delivers the messages to > +the application sent by the other applications in the cluster through the > MDS. > + > + > +CONFIGURATION > + > +Implementation defined configuration is done using a configuration > +file with environment variables. This config file is typically > `/etc/opensaf/dtmd.conf` > +the configuration options provided in are are only valid when MDS transport > is chosen as TCP. > + > +A tipical Four Node Cluster with TCP Transoprt with Broadcost discovery > Configuration: > +-------------------------------------------------------------------------------------- > + > +Note: This following steps assumes that all node IP are in the same subnet > and node > + IP are reachable to each other (one to all ). > + > +Following are the complete Opensaf broadcast cluster configuration > +along with other regular configurations: > + > +1. Configure `DTM_NODE_IP` with IPV4 or IPV6 addresses of that puticuler > Node in > +`/etc/opensaf/dtmd.conf` > + > +2. Configure `MDS_TRANSPORT` with `TCP` transport protocol > /etc/opensaf/nid.conf` > + > + > +A tipical Four Node Cluster with TCP Transoprt with Multicost discovery > Configuration: > +-------------------------------------------------------------------------------------- > +Note: This following steps assumes that your are configuring with two/more > subnet nodes > +and both/more subnets IP are reachable to each other (one to all ).. > + > +Following are the complete Opensaf multicast cluster configuration along > +with other regular configurations: > + > +1. Configure `DTM_NODE_IP` with `IPV4` or `IPV6` addresses of > + that puticuler Node in `/etc/opensaf/dtmd.conf` > +2. Configure DTM_MCAST_ADDR with multicast IP addresses (say 224.0.0.1). > +3. Add DTM_MCAST_ADDR configured multicast address to specific > + interface in all Controller ,Payload Nodes which : > + > + Example: > + SC-1#ip route add 224.0.0.1/4 dev eth0 > + SC-2#ip route add 224.0.0.1/4 dev eth0 > + PL-3#ip route add 224.0.0.1/4 dev eth0 > + PL-4#ip route add 224.0.0.1/4 dev eth0 > + > +4. Enable packet forwarding on all the Controller ,Payload Nodes > + Example: > + > + SC-1#echo 1 > /proc/sys/net/ipv4/ip_forward > + SC-2#echo 1 > /proc/sys/net/ipv4/ip_forward > + PL-3#echo 1 > /proc/sys/net/ipv4/ip_forward > + PL-4#echo 1 > /proc/sys/net/ipv4/ip_forward > + > +Note: In case you don`t have Router for quickly test the above configuration > in your > +Lab , configure a Linux system as a virtual Gateway Router as follows and > verify the > +configuration in your lab. > + > +Continue the below Steps to configure Linux system as a virtual Gateway > Router > +in addition to above steps: > + > +5. configure A simple Linux system with two interfaces > + > + Example: > + Linux system # eth0: 1.1.22.254 > + Linux system # eth1: 1.1.44.254 > + > + > +6. Configure the routes on both the controllers and payloads Nodes > + > + Example: > + SC-1#route add -net 1.1.22.0 netmask 255.255.255.0 gw 1.1.22.254 > + SC-2#route add -net 1.1.44.0 netmask 255.255.255.0 gw 1.1.22.254 > + PL-3#route add -net 1.1.22.0 netmask 255.255.255.0 gw 1.1.44.254 > + PL-4#route add -net 1.1.44.0 netmask 255.255.255.0 gw 1.1.44.254 > + > + > +7. Enable packet forwarding on also on Virtual Gateway Router Linux system > + as mentioned above step 4 for Controller, Payload Nodes: > + > + Example: > + Linux-system# echo 1 > /proc/sys/net/ipv4/ip_forward > + > +8. After this configuration all Controller ,Payload Nodes should be able to > ping > + each other you can verify as follows : > + > + Example: > + SC-1# ping 1.1.22.32 > + SC-1# ping 1.1.44.51 > + SC-1# ping 1.1.44.52 > + > + SC-2# ping 1.1.22.31 > + SC-2# ping 1.1.44.51 > + SC-2# ping 1.1.44.52 > + > + PL-3# ping 1.1.22.31 > + PL-3# ping 1.1.22.32 > + PL-3# ping 1.1.44.52 > + > + PL-4# ping 1.1.22.31 > + PL-4# ping 1.1.22.32 > + PL-4# ping 1.1.44.51 > + > +9. Now,start controllers and payloads with the configurations suggested in > emails > + and should be able from a 4 node cluster . > + > + > +DEBUG > + > +dtm server traces are by default disabled. To enable/disable ntf server > traces > +in a running system, send signal USR2 to the dtmd process. Every time the > +trace state is toggled. Example: > + > + $ pkill -USR2 osafdtmd > + > +Traces are written to the file: > + > + $pklogdir/osafdtmd.log > + > +To enable traces from the very start of the dtm server, uncomment the line: > + > + #args="--tracemask=0xffffffff" > + > +in dtmd.conf (see CONFIGURATION above) and restart the cluster. > + > +For fatal errors syslog is used. > + > > ------------------------------------------------------------------------------ > BPM Camp - Free Virtual Workshop May 6th at 10am PDT/1PM EDT > Develop your own process in accordance with the BPMN 2 standard > Learn Process modeling best practices with Bonita BPM through live exercises > http://www.bonitasoft.com/be-part-of-it/events/bpm-camp-virtual- event?utm_ > source=Sourceforge_BPM_Camp_5_6_15&utm_medium=email&utm_campaign=VA_SF > _______________________________________________ > Opensaf-devel mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/opensaf-devel > ------------------------------------------------------------------------------ BPM Camp - Free Virtual Workshop May 6th at 10am PDT/1PM EDT Develop your own process in accordance with the BPMN 2 standard Learn Process modeling best practices with Bonita BPM through live exercises http://www.bonitasoft.com/be-part-of-it/events/bpm-camp-virtual- event?utm_ source=Sourceforge_BPM_Camp_5_6_15&utm_medium=email&utm_campaign=VA_SF _______________________________________________ Opensaf-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/opensaf-devel
