<?xml version="1.0" encoding="UTF-8"?>
<rss xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:taxo="http://purl.org/rss/1.0/modules/taxonomy/" version="2.0">
  <channel>
    <title>topic Re: Cluster interconnect and priorisation in Operating System - OpenVMS</title>
    <link>https://community.hpe.com/t5/operating-system-openvms/cluster-interconnect-and-priorisation/m-p/5084129#M87618</link>
    <description>I believe that there will be some SCS packets (broadcasts/announcements) on any network interconnect that has SCA enabled, even if there is no currently active connection to another VMS node on that interconnect.&lt;BR /&gt;&lt;BR /&gt;If so, IMO, it is reasonable to disable SCA on certain network interfaces, such as the interface you might use for TCP/IP external internet traffic for security reasons.</description>
    <pubDate>Thu, 13 Dec 2007 17:58:42 GMT</pubDate>
    <dc:creator>Jess Goodman</dc:creator>
    <dc:date>2007-12-13T17:58:42Z</dc:date>
    <item>
      <title>Cluster interconnect and priorisation</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/cluster-interconnect-and-priorisation/m-p/5084124#M87613</link>
      <description>Hello,&lt;BR /&gt;&lt;BR /&gt;I've got a cluster with 3 nodes (Alpha, OpenVMS 7.3-2) and I want to know, how I can see on what connection is the cluster interconnect (I want to have it on a special device). And if the cluster IC is on the wrong device, how can I change this or priorise this.&lt;BR /&gt;&lt;BR /&gt;Regards,&lt;BR /&gt;&lt;BR /&gt;Kirsten</description>
      <pubDate>Thu, 13 Dec 2007 11:50:46 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/cluster-interconnect-and-priorisation/m-p/5084124#M87613</guid>
      <dc:creator>Kirsten Knüttel</dc:creator>
      <dc:date>2007-12-13T11:50:46Z</dc:date>
    </item>
    <item>
      <title>Re: Cluster interconnect and priorisation</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/cluster-interconnect-and-priorisation/m-p/5084125#M87614</link>
      <description>Hi Kirsten&lt;BR /&gt;&lt;BR /&gt;You can use the SCACP program&lt;BR /&gt;$MCR SCACP&lt;BR /&gt;SHOW vc&lt;BR /&gt;SHOW circ&lt;BR /&gt;SHOW channel&lt;BR /&gt;&lt;BR /&gt;You can also set priorities on the various interconnects via the&lt;BR /&gt;SET CHANNEL/PRIO&lt;BR /&gt; &lt;BR /&gt;Fekko</description>
      <pubDate>Thu, 13 Dec 2007 12:08:03 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/cluster-interconnect-and-priorisation/m-p/5084125#M87614</guid>
      <dc:creator>Fekko Stubbe</dc:creator>
      <dc:date>2007-12-13T12:08:03Z</dc:date>
    </item>
    <item>
      <title>Re: Cluster interconnect and priorisation</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/cluster-interconnect-and-priorisation/m-p/5084126#M87615</link>
      <description>Kirsten,&lt;BR /&gt;&lt;BR /&gt;you could do following:&lt;BR /&gt;&lt;BR /&gt;HPS1:SYSTEM&amp;gt; mc scacp sho lan&lt;BR /&gt;&lt;BR /&gt;SMHPS1 PEA0 Device Summary 13-DEC-2007 13:06:26.67:&lt;BR /&gt;&lt;BR /&gt;         Device   Errors +                             Mgt   Buffer  MgtMax  Line     Total         Current&lt;BR /&gt;  Device  Type     Events    Status                 Priority  Size   BufSiz  Speed  Pkts(S+R)     LAN Address&lt;BR /&gt;  ------  ----     ------    ------                 --------  ----   ------  -----  ---------     -----------&lt;BR /&gt;   LCL                  0  Run Online Local Restart       0   1426       0    N/A      399498  00-00-00-00-00-00&lt;BR /&gt;   EWC                  0  Run Online Restart             0   1426       0   1000     2320236  00-1A-4B-AC-B1-40&lt;BR /&gt;   EWD                  0  Run Online Restart             0   1426       0   1000     2237576  00-1A-4B-AC-B1-41&lt;BR /&gt;   EWA                  0  Not in use by SCA              0      0       0   1000           0  00-00-00-00-00-00&lt;BR /&gt;   EWB                  0  Not in use by SCA              0      0       0   1000           0  00-00-00-00-00-00&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;It will show you interconnects used for Cluster.&lt;BR /&gt;&lt;BR /&gt;In scacp you can set things for devices. in your systartup_vms.com add for instance:&lt;BR /&gt;&lt;BR /&gt;$     IF F$GETDVI("EIC0","EXISTS") THEN MC SCACP SET LAN/PRIO=100 EIC&lt;BR /&gt;$     IF F$GETDVI("EID0","EXISTS") THEN MC SCACP SET LAN/PRIO=100 EID&lt;BR /&gt;$     IF F$GETDVI("EWA0","EXISTS") THEN MC SCACP STOP LAN EWA&lt;BR /&gt;$     IF F$GETDVI("EWB0","EXISTS") THEN MC SCACP STOP LAN EWB&lt;BR /&gt;&lt;BR /&gt;This will disable two devices and give others higher priority.&lt;BR /&gt;&lt;BR /&gt;gr,&lt;BR /&gt;&lt;BR /&gt;Robert</description>
      <pubDate>Thu, 13 Dec 2007 12:10:15 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/cluster-interconnect-and-priorisation/m-p/5084126#M87615</guid>
      <dc:creator>R. Verkerk</dc:creator>
      <dc:date>2007-12-13T12:10:15Z</dc:date>
    </item>
    <item>
      <title>Re: Cluster interconnect and priorisation</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/cluster-interconnect-and-priorisation/m-p/5084127#M87616</link>
      <description>Kirsten,&lt;BR /&gt;&lt;BR /&gt;MC SCACP is your friend ;-)&lt;BR /&gt;&lt;BR /&gt;MC SCACP SHOW CHANNEL (use SET TERM/WID=132) will show you the channels, i.e. all cluster communication pathes formed between all visible LAN interfaces of all nodes within the cluster. There can be multiple channels between any 2 nodes in the cluster, it depends on the HW and network configuration.&lt;BR /&gt;&lt;BR /&gt;Between each pair of nodes in the cluster then exists the VC (Virtual Circuit), which carries the SCS traffic. This virtual circuit may be multiplexed over all available channels between that pair of nodes.&lt;BR /&gt;&lt;BR /&gt;You can use various SCACP&amp;gt; SET {CHANNEL|CIRCUIT|PORT|LAN_DEVICE} /PRIO=n commands to manually try to distribute the load over the various communication pathes.&lt;BR /&gt;&lt;BR /&gt;You could also stop SCS communication over individual LAN interfaces (MC SCACP STOP LAN EWA)&lt;BR /&gt;&lt;BR /&gt;If you want to reduce SCS traffic on a specific LAN segment, you could use MC SCACP SET LAN_DEVICE EWA/PRIO=-10&lt;BR /&gt;&lt;BR /&gt;Volker.</description>
      <pubDate>Thu, 13 Dec 2007 12:11:39 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/cluster-interconnect-and-priorisation/m-p/5084127#M87616</guid>
      <dc:creator>Volker Halle</dc:creator>
      <dc:date>2007-12-13T12:11:39Z</dc:date>
    </item>
    <item>
      <title>Re: Cluster interconnect and priorisation</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/cluster-interconnect-and-priorisation/m-p/5084128#M87617</link>
      <description>Kirsten,&lt;BR /&gt;&lt;BR /&gt;Robert wrote:&lt;BR /&gt;&lt;BR /&gt;&amp;gt;&amp;gt;&amp;gt;$&lt;BR /&gt; IF F$GETDVI("EWA0","EXISTS") THEN MC SCACP STOP LAN EWA&lt;BR /&gt;$ IF F$GETDVI("EWB0","EXISTS") THEN MC SCACP STOP LAN EWB&lt;BR /&gt;&lt;BR /&gt;This will disable two devices and give others higher priority.&lt;BR /&gt;&amp;lt;&amp;lt;&amp;lt;&lt;BR /&gt;&lt;BR /&gt;While this is true, you should VERY SERIOUSLY think whether this is what you want/need!&lt;BR /&gt;&lt;BR /&gt;See Volker's answer about LOWERING priority. SHOULD anything unwelcome happen then, then the LOW priority interconnect will keep your cluster alive, but DISABLED does what it says: they will simply NOT function. And if there is NO (enabled) interconnect, there is NO cluster... It will hang, then crash.&lt;BR /&gt;&lt;BR /&gt;Usually a low-performing cluster is WAY preferable, and upon return of the desired interconnect, full function is IMMEDIATELY, AUTOMATICALLY restored.&lt;BR /&gt;&lt;BR /&gt;hth&lt;BR /&gt;&lt;BR /&gt;Proost.&lt;BR /&gt;&lt;BR /&gt;Have one on me.&lt;BR /&gt;&lt;BR /&gt;jpe&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;</description>
      <pubDate>Thu, 13 Dec 2007 12:26:10 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/cluster-interconnect-and-priorisation/m-p/5084128#M87617</guid>
      <dc:creator>Jan van den Ende</dc:creator>
      <dc:date>2007-12-13T12:26:10Z</dc:date>
    </item>
    <item>
      <title>Re: Cluster interconnect and priorisation</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/cluster-interconnect-and-priorisation/m-p/5084129#M87618</link>
      <description>I believe that there will be some SCS packets (broadcasts/announcements) on any network interconnect that has SCA enabled, even if there is no currently active connection to another VMS node on that interconnect.&lt;BR /&gt;&lt;BR /&gt;If so, IMO, it is reasonable to disable SCA on certain network interfaces, such as the interface you might use for TCP/IP external internet traffic for security reasons.</description>
      <pubDate>Thu, 13 Dec 2007 17:58:42 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/cluster-interconnect-and-priorisation/m-p/5084129#M87618</guid>
      <dc:creator>Jess Goodman</dc:creator>
      <dc:date>2007-12-13T17:58:42Z</dc:date>
    </item>
    <item>
      <title>Re: Cluster interconnect and priorisation</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/cluster-interconnect-and-priorisation/m-p/5084130#M87619</link>
      <description>As Jess points out, if you're using a network interconnect (and based on your question, I'm making that assumption), VMS will use all interfaces for cluster traffic.  The operating system will monitor performance and use the interconnect with the best response time by default.  &lt;BR /&gt;&lt;BR /&gt;&amp;gt;&amp;gt;&amp;gt; if the cluster IC is on the wrong device, how can I change this or priorise this.&lt;BR /&gt;&lt;BR /&gt;There should be no "wrong device" as a cluster interconnect.  To use multiple network interfaces as cluster interconnect, simply connect the cables and make sure speed/duplex are properly configured.  If you have multiple interfaces, ideally you'd want to use redundant switching for a 3 node cluster, otherwise the loss of a switch can cause the cluster to fail.  &lt;BR /&gt;&lt;BR /&gt;The SCACP utility will allow you to set priorities and monitor traffic.  Assuming two networks, if you configure one for users and the second for cluster traffic, if you ever have a failure of the cluster network segment, your cluster fails.  If you accept the default behavior however, your cluster would continue.  &lt;BR /&gt;&lt;BR /&gt;</description>
      <pubDate>Thu, 13 Dec 2007 21:16:11 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/cluster-interconnect-and-priorisation/m-p/5084130#M87619</guid>
      <dc:creator>Andy Bustamante</dc:creator>
      <dc:date>2007-12-13T21:16:11Z</dc:date>
    </item>
    <item>
      <title>Re: Cluster interconnect and priorisation</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/cluster-interconnect-and-priorisation/m-p/5084131#M87620</link>
      <description>Thank you to all for your good answers. I know now what to do (and better not to do).&lt;BR /&gt;Thank you very much.&lt;BR /&gt;&lt;BR /&gt;Kirsten</description>
      <pubDate>Sat, 15 Dec 2007 12:02:18 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/cluster-interconnect-and-priorisation/m-p/5084131#M87620</guid>
      <dc:creator>Kirsten Knüttel</dc:creator>
      <dc:date>2007-12-15T12:02:18Z</dc:date>
    </item>
  </channel>
</rss>

