Building a VAX/VMS Cluster
INTRODUCTIONThis project was to build a VAX/VMS cluster using two VAX computers, a VAXStation 3100 (FALCON) and a MicroVAX 3100 Model 40 (HAWK). The goal was to get a local area VAX cluster (LAVC) running, then use it as a basis for adding satellite nodes; the first being a VAX 4000 Model 300 (EAGLE). These computers had not been powered on for over 10 years and the condition of the hardware and disks were unknown. As well, these are systems that were manufactured over 30 years ago and electronic components may have degraded to the point that these systems may be non-functioning.
VMS is a very reliable operating system. Being command line driven, it has a multitude of commands that as a user you need to read up on, remember them and use them often enough so you don't forget them! It is a command rich, complex, operating system that is extremely well documented. This is both a strength and a weakness. Some products are so extensively documented, that it's enough to scare the user away from reading hundreds of pages. However, keep in mind this is a commercial platform and difficult for the occasional hobbyist to tinker with, my opionion. It's not Windows! You need time to read and learn about managing a VMS based system. As with any learning, you have to keep it fresh by using it occasionally so you don't forget it.
HARDWARE CONFIGURATION
| Component | FALCON | HAWK | EAGLE |
|---|---|---|---|
| Model | VAXstation 3100 | MicroVAX 3100/40 | VAX4000/300 |
| CPU | VS42-A | KA45-A | L4000-AA |
| RAM | 24MBytes | 32MBytes | 132MBytes |
| Disks | DKA200 (2.1GB), DKB0 (1.04GB), DKB200 (2.1GB), DKB300 (1.04GB), DKB500 (CD-ROM) | DKA0 (1.04GB), DKA100 (2.1GB), DKA200 (2.1GB), DKA300 (2.1GB), DKA300 (2.1GB), DKA400 (2.1GB) | none |
| Boot Device | DKB0: (master) or ESA0: (satellite) | DKA0: | EZA0: |
| Busses | SCSI A,B | SCSI A | QBus, DSSI A,B |
Both FALCON and HAWK have external storage expansion chassis where additional disk storage is available. One challenge experienced with HAWK was that FALCON's CD-ROM was not recognized when attached to it. To install software on HAWK, FALCON must be booted as a satellite node to make it's CD-ROM available to HAWK over LAVC. Then installation products can be copied from FALCON's CD-ROM over to a hard disk on HAWK, from which product installations can then proceed.
FALCON has two SCSI buses known as A and B bus. Bus A is the internal bus and Bus B is the external SCSI bus attached
to the storage expansion. Each bus can accommodate up to 7 SCSI devices (disks/tapes).
FALCON can theoretically
accommodate a total of 14 disks. I was surprised to find that HAWK, had an external SCSI bus expansion connection,
but it was an extension of its internal A Bus. Subsequently, HAWK can only accommodate 7 disks as it
has only one SCSI bus. On power-up of HAWK's expansion chassis, one of the RZ28 disks started showing an
increasing number of errors as files were being copied to it. Subsequent bootups showed that it just failed
and had to be replaced.
VAX Systems
As a starting point FALCON already had VMS 7.1 installed and running DECNET Phase V. It was also a VAX cluster boot node since it had all the required licensing. HAWK also had VMS 7.1 and Decnet Phase V installed. EAGLE had no operating system disk as it had failed many years ago and therefore was to be used as a satellite node.
The initial thought was to add software to FALCON to make it the cluster master. However, when adding software to FALCON, I noticed that some software such as TCP/IP, MOTIF and some others had more recent releases available than the versions already installed. These newer versions required VMS 7.3. FALCON was already configured with several layered products, but to use the most current versions required an operating system upgrade from 7.1 to 7.3. I expected that the new OS would require more memory, so would any additonal layered products, such as disk shadowing and disk defragmentation products.
Given FALCON's small memroy of 24MBytes as compared to HAWK's 32MBytes, it was not feasible to upgrade FALCON but instead make HAWK the new cluster master. This leaves a configuration whereby FALCON runs VMS 7.1 and can act as a cluster boot system and HAWK runs VMS 7.3 and can also act as a cluster boot system. Only, either HAWK or FALCON can be the boot master at any one time. There can only be one boot master while all other VAX are satellites.
It wasn't possible to use EAGLE as a cluster manager. Given it's larger memory and higher performance (8 VUPs) it would have been the better server. The problem with EAGLE is that its DSSI active terminator circuit is faulty, as a result, the two internal DSSI disks and TK70 tape drive are no longer seen by the CPU; the machine has effectively no local storage. That was a setback since the disks were operational when the machine was first purchased. Such are the challenges with reviving 30 year old systems!
HAWK was therefore chsosen as the 7.3 system and all other satellites would network boot from their respective images on HAWK's system disk. The cluster quorum disk is a local disk on HAWK. The MODPARAMS.DAT file was updated to move the quorum disk from the system disk to a seperate disk, thereby alleviating the system disk of more I/O traffic. Any cluster member is added on HAWK and the satellite boot images are located on HAWK's system disk. Volume shadowing was not used as there is only one SCSI bus on HAWK and the overhead of shadowing two disks on the same bus would have been burdensome for such an underpowered VAX.
EAGLE can still boot from HAWK and join the cluster (Figure 1), but its PAGE and SWAP files being served on HAWK's system disk along with its VMS boot image, reduces the efficacy of this system due to lack of local disks. Because of its large memory, swapping should not be an issue but am not sure if paging will still occur. In either case, all disk I/O happens over a slow 10MBit connection, only alleviated by the fact it's on a switch and not on a HUB.
| Figure 1 |
|---|
>>>b eza0:(BOOT/R5:0 EZA0:) 2..-EZA0 1..0.. %VAXcluster, system loaded from node HAWK (AA-00-04-00-17-04) %SYSBOOT-I-SYSBOOT Mapping the SYSDUMP.DMP on the System Disk%SYSBOOT-W-SYSBOOT Can not map SYSDUMP.DMP on the System Disk%SYSBOOT-I-SYSBOOT Mapping PAGEFILE.SYS on the System Disk %SYSBOOT-I-SYSBOOT SAVEDUMP parameter not set to protect the PAGEFILE.SYS OpenVMS (TM) VAX Version V7.3 Major version id = 1 Minor version id = 0 %DECnet-I-LOADED, network base image loaded, version = 05.0E.00 %DECnet-W-NOOPEN, could not open SYS$SYSROOT:[SYSEXE]NET$CONFIG.DAT %WBM-I-WBMINFO Write Bitmap has successfully completed initialization. %CNXMAN, using remote access method for quorum disk Please enter date and time (DD-MMM-YYYY HH:MM): 05-MAR-2023 10:41 %SYSINIT, waiting to form or join a VMScluster system %CNXMAN, sending VAXcluster membership request to system HAWK %CNXMAN, now a VAXcluster member -- system EAGLE %CNXMAN, established "connection" to quorum disk |
The installation of VMS 7.3 will not be listed here. It is a multi-step installation and is time consuming. The installation guide is well written and easy to follow the various steps. I recall that an ANALYZE/DISK needs to be performed on disks before the installation. Needless to say that the VMS installation went well and layered products were added thereafter.
Below is a display of some of the processes running on HAWK.
HAWK running processes
Below is a display of HAWK's memory utilization.
HAWK memory allocation
Parameter changes can be made on each node through the MODPARAMS.DAT file. Each cluster participant (master and satellites) will have their own MODPARAMS file. In this case the MODPARAMS file for SYS0 (HAWK), was modified to have the quorum disk on DKA100 instead of the system disk. Once parameters are changed, the AUTOGEN program should be run to effect the changes. Be sure to check the resulting report (Figure 6) for the changes made, overrides or errors encountered. It is very informative.
MODPARAMS.DAT file locations
MODPARAMS.DAT sample file
Running AUTOGEN
The steps below create new page and swap files on DKA200. This should alleviate the load on the system disk by eventually adding these files to the pool of page and swap files. These files can reside on more than one disk simultaneously; it does not have to be one or the other.
Examine current page/swap sizes
Creation and installation of page/swap files
Several layered products were installed on HAWK (see below). DECWindows/MOTIF was installed, giving HAWK, the ability to display an X-Window environment on an Apple laptop running OSX 10. I prefer to use Apple OSX for X-Windows because they are relatively inexpensive system to acquire and the DEC terminal emulation is better than Putty on Windows. The functions keys on the LK201 keyboard are decently emulated on OSX's X-Terminal.
Active licenses on node HAWK: ------- Product ID -------- Product Producer C DEC CDD-PLUS DEC DFG DEC DVNETRTG DEC DW-MOTIF DEC ELMS DEC FORTRAN DEC NET-APP-SUP-200 DEC POLY-SWDAGT DEC POLY-SWDCON-USER DEC VAX-VMS DEC VAXCLUSTER DEC VAXSET DECDECNET V (OSI)
On FALCON I found it difficult to make changes to Decnet V. As well there were some problems from the initial installstion that I could not figure out how to correct. I decided to remove it and install Decnet IV, which is much simpler to configure, manage and is more familiar to me. I suspect it also uses less memory. I found instructions on the internet for remocing Decnet V and installing IV. The process is outlined below:
- Stop Decnet using; STOP/NETWORK DECNET
- Use the PRODUCT SHOW PRODUCT command to get a list of installed products.
- then PRODUCT REMOVE "product name"
- Reboot the system to be sure all components are removed.
- DECnet Phase IV is a POLYCENTER Software Installation utility kit. It is on the OpenVMS installation system CD-ROM.
- Set directory, [DECNET_PHASE_IV+VAX071.KIT].
- Mount the Decnet IV save sets CD (assuming CD-ROM is on DKA400 as an example).
- $PRODUCT INSTALL DECNET_PHASE_IV/SOURCE=DKA400:[DECNET_PHASE_IV+VAX071.KIT]
- Reboot the system to start up VMS with Decnet IV.
Use NCP (Network Control Program) to clear existing node settings. These same steps were followed on FALCON and HAWK.
- $ run sys$system:ncp
- NCP> clear node *
- NCP> purge node *
- NCP> exit
Next step, add the Decnet node information into the NCP database.
- $ run sys$system:ncp
- NCP>define node 1.20 name FALCON
- NCP>define node 1.23 name HAWK
- NCP>define node 1.33 name EAGLE
- NCP>define node 1.34 name PRO350
- NCP>set known nodes all
- NCP>show node *
On HAWK, VMS 7.3 was installed with Decnet IV to avoid the complexity issues experienced on FALCON.
HUB/SWITCH PROBLEMSAll VAX nodes are connected to a network switch using 10MBit transceivers manufactured by Allied Telesis (CentreCOM 210TS Twisted Pair transceivers). The computers are successfully connected to an Intel 10/100 switch. The VAX systems only communicate over 10MBit ethernet connections. However, before landing on the Intel switch, other 10MBit switches were tried without success. It seems that when the transceiver was connected (over an ethernet cable) to some hubs the connection was not recognized. The hub speed indicator, normally, green would not light up at all when connecting the computer and hub.
Two seperate DLINK DSA-8H hubs failed to recognize the connection. Another 16 port 10/100 DLINK hub also failed to establish a connection. A TRENDNET Gigabit switch failed as well. An older HP 28688B HUB Plus recognized the connections but resulted in dropped packets during cluster booting of EAGLE (Figure 2). For anyone attempting to network older VAX systems there could be issues when using more modern network equipment (switches, routers, etc.).
A ThinNet (coaxial) network was considered since the coax connector is available on the VAXes and I also have ThinNet transceivers. However, given that these network configurations are prone to packet collisions as more nodes are added to the network, it didn't make sense to use. ThinNet is one of the simplest to setup on LAVC networks but these networks are prone to network congestion. Network switches far outperform hubs and ThinNet networks.
| Figure 2 |
|---|
%PAA0, Path #0. Loopback has gone from GOOD to BAD %PAA0, Path #0. Loopback has gone from BAD to GOOD %NET$CONFIGURE-I-CONFIGCOMPLETED, DECnet-Plus for OpenVMS configuration completed %AUTOGEN-I-BEGIN, GETDATA phase is beginning. %PAA0, Path #0. Loopback has gone from GOOD to BAD %AUTOGEN-I-NEWFILE, A new version of SYS$SYSTEM:PARAMS.DAT has been created. You may wish to purge this file.%AUTOGEN-I-END, GETDATA phase has successfully completed. %AUTOGEN-I-BEGIN, GENPARAMS phase is beginning. %PEA0, Excessive packet losses on LAN Path from EZA to ESA on REMOTE NODE HAWK %PAA0, CI port timeout. %PAA0, Port is Reinitializing ( 41 Retries Left). Check the Error Log. %PEA0, Excessive packet losses on LAN Path from EZA to ESA on REMOTE NODE HAWK %PEA0, Excessive packet losses on LAN Path from EZA to ESA on REMOTE NODE HAWK %PEA0, Excessive packet losses on LAN Path from EZA to ESA on REMOTE NODE HAWK %PEA0, Excessive packet losses on LAN Path from EZA to ESA on REMOTE NODE HAWK |
As all VAX systems succesfully connected through the Intel switch, one further experiment was to see if the PRO350 could connect to HAWK over ethernet based Decnet. The goal was to establish a method to exchange data between the PRO350 and HAWK. The PRO Series runs P/OS 3.2 (based on RSX) with Decnet 2.0 installed. The PRO350 is a PDP-11 desktop computer, based on the same F-11 chipset that the PDP-11/23 computer uses. It is part of the DEC Professional series of desktop computers encompassing models 325, 350 and 380. Memory can be expanded to 4MBytes, although this system has 256KBytes. More information on the Pro series is on this site.
I was able to test the Decnet connection from the PRO to HAWK using the Decnet diagnostic menu on the PRO. Therefore I was confident that there were no network issues. The problem was finding the correct syntax for moving files from the PRO to the VAX. On the Internet, I located the syntax to exchange files between HAWK and PRO350 if the request originated on HAWK. The screen sample below shows file directory listing being requested from the PRO from HAWK over Decnet.
HAWK directory listing of PRO350 disk
In summary, the project was successful in setting up a VAX cluster system that additional VAX systems can be added to. It was also possible to establish a method to exchange files between the VAX and the PRO350 computer. The limits on HAWK is its small memory size and that as more satellites are added, there are more demands on the system disk and the slow 10MBit network. The small memory size also limits how many applications or layered products can be run on HAWK. If EAGLE is to be used more effectively, then the PAGE and SWAP files that are now served by HAWK over the slow network will ideally have to be moved to local disks, a challenge since the active bus terminator is faulty, resulting in local disks being inaccesible.
On the hardware size there were less challenges than I expected. Other than the faulty EAGLE termintator and a faulty disk on HAWK, all else worked. The VAX systems are well built and after many years, and on/off power cycles, they continue to operate reliably.
ADDENDUMA1 - FALCON Satellite Node
Booting FALCON from HAWK as a satellite node. Note that licenses were not yet configured on FALCON. As well, some packet losses are experinced as the systems are connected throught an HP HUB at this time which was eventually replaced with an Intel switch.
| Figure 3 |
|---|
>>> b esa0:-ESA0 %VAXcluster, system loaded from node HAWK (AA-00-04-00-17-04) %SYSBOOT-I-SYSBOOT Mapping the SYSDUMP.DMP on the System Disk %SYSBOOT-I-SYSBOOT SYSDUMP.DMP on System Disk successfully mapped %SYSBOOT-I-SYSBOOT Mapping PAGEFILE.SYS on the System Disk %SYSBOOT-I-SYSBOOT SAVEDUMP parameter not set to protect the PAGEFILE.SYS OpenVMS (TM) VAX Version V7.3 Major version id = 1 Minor version id = 0 %DECnet-I-LOADED, network base image loaded, version = 05.0C.00 %DECnet-W-NOOPEN, could not open SYS$SYSROOT:[SYSEXE]NET$CONFIG.DAT %WBM-I-WBMINFO Write Bitmap has successfully completed initialization. %CNXMAN, using remote access method for quorum diskPLEASE ENTER DATE AND TIME (DD-MMM-YYYY HH:MM) D-MMM-YYYY HH:MM) 03-MAR-2023 15:14 %SYSINIT, waiting to form or join a VMScluster system %CNXMAN, have connection to system HAWK %CNXMAN, sending VAXcluster membership request to system HAWK %CNXMAN, now a VAXcluster member -- system FALCON %CNXMAN, quorum lost, blocking activity %CNXMAN, established "connection" to quorum disk %CNXMAN, quorum regained, resuming activity %SYSINIT, primary PAGEFILE.SYS not found; system initialization continuing %MSCPLOAD-I-LOADMSCP, loading the MSCP disk server %PEA0, Excessive packet losses on LAN Path from ESA to ESA on REMOTE NODE HAWK %PEA0, Excessive packet losses on LAN Path from ESA to ESA on REMOTE NODE HAWK %MSCPLOAD-I-CONFIGSCAN, enabled automatic disk serving ***** Mounting page/swap device: FALCON$DKB300 %MOUNT-I-MOUNTED, FALCON_1044 mounted on _FALCON$DKB300: %PEA0, Excessive packet losses on LAN Path from ESA to ESA on REMOTE NODE HAWK RUN-S-PROC_ID, identification of created process is 20400107On MIN or UPGRADE system startup - CLUE is not run. %PEA0, Excessive packet losses on LAN Path from ESA to ESA on REMOTE NODE HAWK %%%%%%%%%%% OPCOM 20-FEB-2023 12:34:07.41 %%%%%%%%%%% Operator _FALCON$OPA0: has been enabled, username SYSTEM %%%%%%%%%%% OPCOM 20-FEB-2023 12:34:07.46 %%%%%%%%%%% Operator status for operator _FALCON $OPA0:CENTRAL, PRINTER, TAPES, DISKS, DEVICES, CARDS, NETWORK, CLUSTER, SECURITY,LICENSE, OPER1, OPER2, OPER3, OPER4, OPER5, OPER6, OPER7, OPER8, OPER9, OPER10,OPER11, OPER12 %%%%%%%%%%% OPCOM 20-FEB-2023 12:34:07.86 %%%%%%%%%%% Logfile has been initialized by operator _FALCON $OPA0:Logfile is FALCON::SYS$SYSROOT:[SYSMGR]OPERATOR.LOG;3 %%%%%%%%%%% OPCOM 20-FEB-2023 12:34:07.87 %%%%%%%%%%% Operator status for operator FALCON::SYS$SYSROOT:[SYSMGR]OPERATOR.LOG; CENTRAL, PRINTER, TAPES, DISKS, DEVICES, CARDS, NETWORK, CLUSTER, SECURITY,LICENSE, OPER1, OPER2, OPER3, OPER4, OPER5, OPER6, OPER7, OPER8, OPER9, OPER10,OPER11, OPER12 %SYSTEM-I-BOOTUPGRADE, security auditing disabled %%%%%%%%%%% OPCOM 20-FEB-2023 12:34:08.08 %%%%%%%%%%% 00:00:05.07 Node FALCON (csid 00000000) is using remote access method for quorum disk %%%%%%%%%%% OPCOM 20-FEB-2023 12:34:08.09 %%%%%%%%%%% 15:14:55.00 Node FALCON (csid 00000000) has connection to node HAWK %%%%%%%%%%% OPCOM 20-FEB-2023 12:34:08.10 %%%%%%%%%%% 15:14:55.44 Node FALCON (csid 00010002) is now a VAXcluster member %%%%%%%%%%% OPCOM 20-FEB-2023 12:34:08.12 %%%%%%%%%%% 15:15:00.05 Node FALCON (csid 00010002) lost quorum, blocking activity %%%%%%%%%%% OPCOM 20-FEB-2023 12:34:08.14 %%%%%%%%%%% 15:15:05.04 Node FALCON (csid 00010002) re-established "connection" to quorum disk %%%%%%%%%%% OPCOM 20-FEB-2023 12:34:08.15 %%%%%%%%%%% 15:15:05.05 Node FALCON (csid 00010002) regained quorum, proceeding %%%%%%%%%%% OPCOM 20-FEB-2023 12:34:08.18 %%%%%%%%%%% Operator _HAWK$OPA0: has been enabled, username SYSTEM %LICENSE-W-NOEXCL, EXCLUDE list not permitted with NO_SHARE option; use an INCLUDE list for BASE-VMS-250136 to be loaded %%%%%%%%%%% OPCOM 20-FEB-2023 12:34:11.95 %%%%%%%%%%% Message from user JOB_CONTROL on FALCON %JBC-E-OPENERR, error opening SYS$COMMON:[SYSEXE]QMAN$MASTER.DAT; %%%%%%%%%%% OPCOM 20-FEB-2023 12:34:11.96 %%%%%%%%%%% Message from user JOB_CONTROL on FALCON-RMS-E-FNF, file not found %LICENSE-W-NOLOAD, license was not loaded for DVNETRTG-LICENSE-F-EXCEEDED, attempted usage exceeds active license limits %%%%%%%%%%% OPCOM 20-FEB-2023 12:34:17.90 %%%%%%%%%%% Message from user SYSTEM on FALCON%LICENSE-W-NOLOAD, license was not loaded for DVNETRTG %LICENSE-W-NOLOAD, license was not loaded for ELMS %%%%%%%%%%% OPCOM 20-FEB-2023 12:34:17.93 %%%%%%%%%%% Message from user SYSTEM on FALCON%LICENSE-F-EXCEEDED, attempted usage exceeds active license limits -LICENSE-F-EXCEEDED, attempted usage exceeds active license limit %%%%%%%%%%% OPCOM 20-FEB-2023 12:34:20.33 %%%%%%%%%%% Message from user SYSTEM on FALCON%LICENSE-W-NOLOAD, license was not loaded for ELMS %%%%%%%%%%% OPCOM 20-FEB-2023 12:34:20.35 %%%%%%%%%%% Message from user SYSTEM on FALCON%LICENSE-F-EXCEEDED, attempted usage exceeds active license limits %SYSTEM-I-BOOTUPGRADE, security server not started %%%%%%%%%%% OPCOM 20-FEB-2023 12:34:22.55 %%%%%%%%%%% (from node HAWK at 20-FEB-2023 12:34:22.54) Message from user SYSTEM on HAWKEvent: Late Collision from: Node LOCAL:.HAWK CSMA-CD Station CSMACD-0, at: 2023-02-20-12:34:22.530-05:00Iinf eventUid E73F3A20-B11A-11ED-A093-08002B1B5959 entityUid 3E53CF80-B047-11ED-8002-AA0004001704 streamUid 50093260-B047-11ED-8003-AA0004001704 NET$STARTUP, Network not started due to UPGRADE boot %%%%%%%%%%% OPCOM 20-FEB-2023 12:34:26.60 %%%%%%%%%%% (from node HAWK at 20-FEB-2023 12:34:26.77) Message from user SYSTEM on HAWKEvent: Late Collision from: Node LOCAL:.HAWK CSMA-CD Station CSMACD-0, at: 2023-02-20-12:34:26.750-05:00Iinf eventUid E9C325E0-B11A-11ED-A093-08002B1B5959 entityUid 3E53CF80-B047-11ED-8002-AA0004001704 streamUid 50093260-B047-11ED-8003-AA0004001704 %%%%%%%%%%% OPCOM 20-FEB-2023 12:34:26.86 %%%%%%%%%%% (from node HAWK at 20-FEB-2023 12:34:27.13) Message from user SYSTEM on HAWKEvent: Late Collision from: Node LOCAL:.HAWK CSMA-CD Station CSMACD-0, at: 2023-02-20-12:34:27.110-05:00Iinf eventUid E9FA1460-B11A-11ED-A093-08002B1B5959 entityUid 3E53CF80-B047-11ED-8002-AA0004001704 streamUid 50093260-B047-11ED-8003-AA0004001704 %%%%%%%%%%% OPCOM 20-FEB-2023 12:34:30.04 %%%%%%%%%%% Message from user SYSTEM on FALCON%LICENSE-E-NOAUTH, DEC VAX-VMS use is not authorized on this node -LICENSE-F-NOLICENSE, no license is active for this software product -LICENSE-I-SYSMGR, please see your system manager %LICENSE-E-NOAUTH, DEC VAX-VMS use is not authorized on this node-LICENSE-F-NOLICENSE, no license is active for this software product-LICENSE-I-SYSMGR, please see your system managerStartup processing continuing... %MSCPLOAD-I-CONFIGSCAN, enabled automatic disk serving ********************************************************************* A new SYS$SPECIFIC:[SYSEXE]MODPARAMS.DAT file was created for this node during the upgrade and will now be used. See MODPARAMS.DAT for more information on how this file was built. ********************************************************************* Running AUTOGEN ...%AUTOGEN-I-BEGIN, GETDATA phase is beginning. %AUTOGEN-I-NEWFILE, A new version of SYS$SYSTEM:PARAMS.DAT has been created. You may wish to purge this file. %AUTOGEN-I-END, GETDATA phase has successfully completed. %AUTOGEN-I-BEGIN, GENPARAMS phase is beginning %%%%%%%%%%% OPCOM 20-FEB-2023 12:35:08.04 %%%%%%%%%%% (from node HAWK at 20-FEB-2023 12:35:08.19)Message from user SYSTEM on HAWK Event: Late Collision from: Node LOCAL:.HAWK CSMA-CD Station CSMACD-0, at: 2023-02-20-12:35:08.170-05:00Iinf eventUid 027356A0-B11B-11ED-A093-08002B1B5959 entityUid 3E53CF80-B047-11ED-8002-AA0004001704 streamUid 50093260-B047-11ED-8003-AA0004001704 |
A2 - Adding EAGLE to the cluster
This section lists the log for adding EAGLE to the cluster as a satellite node. In this scenario HAWK was the cluster boot node and will host the swap and page files. At this time HAWK was still running Decnet V.
| Figure 4 |
|---|
Welcome to OpenVMS (TM) VAX Operating System, Version V7.3 Username: SYSTEM Password: xxxxxx Welcome to OpenVMS (TM) VAX Operating System, Version V7.3 on node HAWK Last interactive login on Tuesday, 21-FEB-2023 13:29 Last non-interactive login on Tuesday, 14-FEB-2023 20:19 Hawk> @SYS$MANAGER:CLUSTER_CONFIG.COM Cluster Configuration Procedure Executing on a VAX System DECnet-Plus is installed on this node. This procedure has detected that the LANACP LAN server process is running on this node. NOTE: LANACP is the most efficient way to provide down line load service to satellites and it is strongly recommended that LANACP be used instead of DECnet. Answer YES to the following question if you want to use LANACP for down line loading satellites. Do you want LANACP instead of DECnet for cluster boot serving? [YES] Cluster Configuration Procedure Executing on a VAX System DECnet-Plus is installed on this node. The LAN, not DECnet, will be used for MOP downline loading. To ensure that this procedure is executing with the required privileges, invoke it from the system manager's account. Enter a "?" for help at any prompt. If you are familiar with the execution of this procedure, you may want to mute extra notes and explanations by invoking it with "@CLUSTER_CONFIG_LAN BRIEF". HAWK is a VAX system and currently a member of a cluster so the following functions can be performed: MAIN MENU 1. ADD a VAX node to the cluster. 2. REMOVE a node from the cluster. 3. CHANGE a cluster member's characteristics. 4. CREATE a duplicate system disk for HAWK. 5. MAKE a directory structure for a new root on a system disk. 6. DELETE a root from a system disk. 7. EXIT from this procedure.Enter choice [1]: This ADD function will add a new VAX node to the cluster. WARNING: If the node being added is a voting member, EXPECTED_VOTES for every cluster member must be adjusted. For complete instructions check the section on configuring a cluster in the "OpenVMS Cluster Systems" manual. CAUTION: If this cluster is running with multiple system disks and common system files will be used, please, do not proceed unless appropriate logical names are defined for cluster common files in SYLOGICALS.COM. For instructions, refer to the "OpenVMS Cluster Systems" manual. Do you want to continue [N]? Y Will the node be a satellite [Y]? Verifying LAN adapters in LANACP database... What is the node's SCS node name? EAGLE DECnet is running on this node. Even though you are configuring a LAN- based cluster, the DECnet database will provide some information and may be updated.What is the node's DECnet fullname? ? A fullname should be determined with the help of your network manager. Enter a string comprised of: - (optional) the namespace name, ending with a colon ":" - the root directory, designated by a period "." - zero or more hierarchical directories, designated by a character string followed by a period - the simple name, a character string that, combined with the directory names, uniquely identifies the node Examples: .SALES.NETWORKS.MYNODE MEGA:.INDIANA.JONES COLUMBUS:.FLATWORLD What is the node's DECnet fullname? LOCAL:.EAGLE Do you want to define a DECnet synonym [Y]? What is the synonym name for this node [EAGLE]? What is the MOP service client name for this node [EAGLE]? What is the node's SCSSYSTEMID number? 1.33 NOTE: 1.33 equates to an SCSSYSTEMID of 1057 WARNING: The DECnet databases on HAWK will not be updated with information on EAGLE. You must see to it that network databases on this and all other cluster members are updated. For help, refer to the "OpenVMS Cluster Systems" manual. Does EAGLE need to be registered in the namespace [N]? Verifying LANACP satellite node database... What is the Cluster Alias fullname? ? The Alias name is the node name of the alias: the DNS FullName of the object which stores the address towers for the alias. Do not enter a Node Synonym. If this node will not be participating in a cluster alias, enter a carriage return. The Cluster Alias name should be determined with the help of your network manager. Enter a string comprised of: - (optional) the namespace name, ending with a colon ":" - the root directory, designated by a period "." - zero or more hierarchical directories, designated by a character string followed by a period - the simple name, a character string that, combined with the directory names, uniquely identifies the node Examples: .NETWORKS.FAROUT MEGA:.PROTO.QUIK KWhat is the Cluster Alias fullname? LOCAL: ERROR: LOCAL: is an invalid node name. The alias must be a DNS fullname. Type ? for help.What is the Cluster Alias fullname? ? The Alias name is the node name of the alias: the DNS FullName of the object which stores the address towers for the alias. Do not enter a Node Synonym. If this node will not be participating in a cluster alias, enter a carriage return. The Cluster Alias name should be determined with the help of your network manager. Enter a string comprised of: - (optional) the namespace name, ending with a colon ":" - the root directory, designated by a period "." - zero or more hierarchical directories, designated by a character string followed by a period - the simple name, a character string that, combined with the directory names, uniquely identifies the node Examples: .NETWORKS.FAROUT MEGA:.PROTO.QUIKKWhat is the Cluster Alias fullname? UGENIX: ERROR: UGENIX: is an invalid node name. The alias must be a DNS fullname. Type ? for help. What is the Cluster Alias fullname? LOCAL:.EAGLE What is the Phase IV address of the Cluster Alias? 1.33 What selection weight do you choose for this node? [0 for satellites] 0 Will EAGLE run DECnet [Y]? This procedure will establish Ethernet adapter ESA0 for HAWK's booting of EAGLE via the LANACP boot server process. What is EAGLE's LAN adapter hardware address? 08-00-2B-27-5B-D1 This procedure will now ask you for the device name of EAGLE's system root. The default device name (DISK$OPENVMS071:) is the logical volume name of SYS$SYSDEVICE:. What is the device name for EAGLE's system root[default DISK$OPENVMS071:]? What is the name of EAGLE's system root [SYS10]? Allow conversational bootstraps on EAGLE [N]? The following workstation windowing options are available: 1. No workstation software 2. DECwindows Workstation Software 3. VWS Workstation Software (not installed)Enter choice [1]: 1 Creating directory tree SYS10 ... System root SYS10 createdWill EAGLE be a disk server [N]? Y Will EAGLE serve RFxx disks [N]? Y Any RFxx disks will be served if they are visible to EAGLE and have the same ALLOCLASS value as EAGLE. Enter a value for EAGLE's ALLOCLASS parameter [0]: Updating LANACP LAN server process volatile and permanent databases... Size of pagefile for EAGLE [RETURN for AUTOGEN sizing]? A temporary pagefile will be created until resizing by AUTOGEN. The default size below is arbitrary and may or may not be appropriate. Size of temporary pagefile [10000]? Size of swap file for EAGLE [RETURN for AUTOGEN sizing]? A temporary swap file will be created until resizing by AUTOGEN. The default size below is arbitrary and may or may not be appropriate. Size of temporary swap file [8000]? NOTE: If an RFxx disk visible to the node will be used for local page and swap files, the disk and node must have the same allocation class. Will a local disk on EAGLE be used for paging and swapping (Y/N)? N If you specify a device other than DISK$OPENVMS071: for EAGLE's page and swap files, this procedure will create PAGEFILE_EAGLE.SYS and SWAPFILE_EAGLE.SYS in the |
A3 - Removing DODO from the cluster
This is the sequence followed for removing a satellite node from the cluster. At one time, the cluster was configured with an incorrect node name; DODO instead of EAGLE. At this time HAWK was still running Decnet V.
| Figure 5 |
|---|
Hawk> @sys$manager:cluster_config.com Cluster Configuration Procedure Executing on a VAX System This system is running DECnet-Plus. DECnet will be used for MOP downline loading. To ensure that this procedure is executing with the required privileges, invoke it from the system manager's account. Enter a "?" for help at any prompt. If you are familiar with the execution of this procedure, you may want to mute extra notes and explanations by invoking it with "@CLUSTER_CONFIG BRIEF". HAWK is a VAX system and currently a member of a cluster so the following functions can be performed: MAIN MENU 1. ADD a VAX node to the cluster. 2. REMOVE a node from the cluster. 3. CHANGE a cluster member's characteristics. 4. CREATE a duplicate system disk for HAWK. 5. MAKE a directory structure for a new root on a system disk. 6. DELETE a root from a system disk. 7. EXIT from this procedure. Enter choice [1]: 2 The REMOVE command disables a node as a cluster member. o It deletes the node's root directory tree. o It removes the node's network information from the network database. o If the node has entries in SYS$DEVICES.DAT, any port allocation class for shared SCSI bus access on the node must be re-assigned. If the node being removed is a voting member, EXPECTED_VOTES in each remaining cluster member's MODPARAMS.DAT must be adjusted. The cluster must then be rebooted. For instructions, see the "OpenVMS Cluster Systems" manual. For additional networking information, please refer to the DECnet-Plus Network Management manual. What is the node's DECnet fullname? ugenix.dodo What is the SCS node name for this node [DODO]? Was a DECnet synonym defined for this node [Y]? y What is the synonym name for this node [DODO]? What is the MOP service client name for this node [DODO]? Are you sure you want to remove node DODO (Y/N)? y Verifying that SYS12 is DODO's root... Deleting directory tree SYS12... %DELETE-I-FILDEL, HAWK$DKA0: |
A4 - AGEN$PARAMS.REPORT
The AGEN$PARAMS.REPORT report is created after AUTOGEN completes. Typically there are a few warning to look into and correct for the next AUTOGEN run. Below is a sample report.
| Figure 6 |
|---|
AUTOGEN Parameter Calculation Report on node: HAWK This information was generated at 25-MAR-2023 07:22:51.06 AUTOGEN was run from GENPARAMS to GENPARAMS - default execution specified ** No changes will be done by AUTOGEN ** The values given in this report are what AUTOGEN would have set the parameters to. The following was detected within MODPARAMS.DAT Please review immediately. ** INFORMATIONAL ** - Duplicate value for VAXCLUSTER found. This overrides preceding requirements for this parameter in PARAMS.DAT Review the parameter setting and if possible use MIN, MAX, or ADD. ** INFORMATIONAL ** - Duplicate values for MIN_CHANNELCNT found. Records were defined by MODPARAMS and VMS to set symbol to 255 ** INFORMATIONAL ** - Multiple MIN values found for MIN_GBLPAGES. Using VMS value (62000) to supersede MODPARAMS value (30000) ** INFORMATIONAL ** - Multiple MIN values found for MIN_GBLSECTIONS. Using MODPARAMS value (520) to supersede VMS value (420) ** WARNING ** - ENQLM is not a SYSGEN parameter: MIN_ENQLM=2048 ** INFORMATIONAL ** - Duplicate value for VAXCLUSTER found. This overrides preceding requirements for this parameter in PARAMS.DAT Review the parameter setting and if possible use MIN, MAX, or ADD. ** INFORMATIONAL ** - Duplicate value for QDSKVOTES found. This overrides preceding requirements for this parameter in PARAMS.DAT Review the parameter setting and if possible use MIN, MAX, or ADD. Feedback information was collected on 15-MAR-2023 11:39:18.95 Old values below are the parameter values at the time of collection. The feedback data is based on 25 hours of up time. Feedback information will be used in the subsequent calculations Parameter information follows: ------------------------------ MAXPROCESSCNT parameter information: Feedback information. Old value was 240, New value is 200 Maximum Observed Processes: 25 Information on VMS executable image Processing: Processing SYS$MANAGER:VMS$IMAGES_MASTER.DAT GBLPAGFIL parameter information: Override Information - parameter calculation has been overridden. The calculated value was 1024. The new value is 6024. GBLPAGFIL has been increased by 5000. GBLPAGFIL is not allowed to be less than 6024. GBLPAGES parameter information: Feedback information. Old value was 78000, New value is 62400 Peak used GBLPAGES: 17128 Global buffer requirements: 6024 GBLSECTIONS parameter information: Feedback information. Old value was 530, New value is 520 Peak used GBLSECTIONS: 254 Override Information - parameter calculation has been overridden. The calculated value was 420. The new value is 520. GBLSECTIONS is not allowed to be less than 520. LOCKIDTBL parameter information: Feedback information. Old value was 2590, New value is 2072 Current number of locks: 1282 Peak number of locks: 2590 LOCKIDTBL_MAX parameter information: Feedback information. Old value was 94152, New value is 16776959 RESHASHTBL parameter information: Feedback information. Old value was 2048, New value is 1024 Current number of resources: 983 MSCP_BUFFER parameter information: Feedback information. Old value was 128, New value is 128 MSCP server I/O rate: 1 I/Os per 10 sec. I/Os that waited for buffer space: 0 I/Os that fragmented into multiple transfers: 18 SCSCONNCNT parameter information: Feedback information. Old value was 12, New value is 12 Peak number of nodes: 2 Number of CDT allocation failures: 0 SCSRESPCNT parameter information: Feedback information. Old value was 300, New value is 300 RDT stall count: 0 SCSBUFFCNT parameter information: Feedback information. Old value was 512, New value is 512 CIBDT stall count: 0 NPAGEDYN parameter information: Feedback information. Old value was 2409984, New value is 2169344 Maximum observed non-paged pool size: 2409984 bytes. Non-paged pool request rate: 0 requests per 10 sec. LNMSHASHTBL parameter information: Feedback information. Old value was 128, New value is 128 Current number of shareable logical names: 226 ACP_DIRCACHE parameter information: Feedback information. Old value was 240, New value is 240 Hit percentage: 89% Attempt rate: 11 attempts per 10 sec. ACP_DINDXCACHE parameter information: Feedback information. Old value was 60, New value is 60 Hit percentage: 91% Attempt rate: 17 attempts per 10 sec. ACP_HDRCACHE parameter information: Feedback information. Old value was 240, New value is 288 Hit percentage: 74% Attempt rate: 44 attempts per 10 sec. ACP_MAPCACHE parameter information: Feedback information. Old value was 16, New value is 16 Hit percentage: 93% Attempt rate: 0 attempts per 10 sec. PAGEDYN parameter information: Feedback information. Old value was 1814016, New value is 1814016 Current paged pool usage: 423184 bytes. Paged pool request rate: 0 requests per 10 sec. SPTREQ parameter information: Override Information - parameter calculation has been overridden. The calculated value was 4156. The new value is 8000. SPTREQ is not allowed to be less than 8000. SYSMWCNT parameter information: Override Information - parameter calculation has been overridden. The calculated value was 2734. The new value is 2929. SYSMWCNT has been increased by 195. PQL_DPGFLQUOTA parameter information: Override Information - parameter calculation has been overridden. The calculated value was 16400. The new value is 32768. PQL_DPGFLQUOTA is not allowed to be less than 32768. PQL_MPGFLQUOTA has been increased to accomodate the working set requirements of PQL_MWSEXTENT. VAXCLUSTER parameter information: Override Information - parameter calculation has been overridden. The calculated value was 1. The new value is 2. VAXCLUSTER has been set to the hard-coded value of 2. EXPECTED_VOTES parameter information: Override Information - parameter calculation has been overridden. The calculated value was 3. The new value is 1. EXPECTED_VOTES has been set to the hard-coded value of 1. PASTDGBUF parameter information: Override Information - parameter calculation has been overridden. The calculated value was 4. The new value is 16. ERRORLOGBUFFERS parameter information: Override Information - parameter calculation has been overridden. The calculated value was 4. The new value is 14. ERRORLOGBUFFERS is not allowed to be less than 14. SCSSYSTEMID parameter information: Override Information - parameter calculation has been overridden. The calculated value was 0. The new value is 1047. SCSSYSTEMID has been set to the hard-coded value of 1047. PQL_MASTLM parameter information: Override Information - parameter calculation has been overridden. The calculated value was 4. The new value is 100. PQL_MASTLM is not allowed to be less than 100. PQL_MBIOLM parameter information: Override Information - parameter calculation has been overridden. The calculated value was 4. The new value is 100. PQL_MBIOLM is not allowed to be less than 100. PQL_MBYTLM parameter information: Override Information - parameter calculation has been overridden. The calculated value was 1024. The new value is 48000. PQL_MBYTLM is not allowed to be less than 48000. PQL_MDIOLM parameter information: Override Information - parameter calculation has been overridden. The calculated value was 4. The new value is 100. PQL_MDIOLM is not allowed to be less than 100. PQL_MFILLM parameter information: Override Information - parameter calculation has been overridden. The calculated value was 2. The new value is 100. PQL_MFILLM is not allowed to be less than 100. PQL_MPRCLM parameter information: Override Information - parameter calculation has been overridden. The calculated value was 0. The new value is 8. PQL_MPRCLM is not allowed to be less than 8. PQL_MENQLM parameter information: Override Information - parameter calculation has been overridden. The calculated value was 30. The new value is 200. PQL_MENQLM is not allowed to be less than 200. QDSKVOTES parameter information: Override Information - parameter calculation has been overridden. The calculated value was 1. The new value is 2. QDSKVOTES has been set to the hard-coded value of 2. QDSKINTERVAL parameter information: Override Information - parameter calculation has been overridden. The calculated value was 3. The new value is 10. QDSKINTERVAL has been set to the hard-coded value of 10. NISCS_LOAD_PEA0 parameter information: Override Information - parameter calculation has been overridden. The calculated value was 0. The new value is 1. NISCS_LOAD_PEA0 has been set to the hard-coded value of 1. MSCP_SERVE_ALL parameter information: Override Information - parameter calculation has been overridden. The calculated value was 4. The new value is 2. MSCP_SERVE_ALL has been set to the hard-coded value of 2. WINDOW_SYSTEM parameter information: Override Information - parameter calculation has been overridden. The calculated value was 0. The new value is 1. WINDOW_SYSTEM has been set to the hard-coded value of 1. CHANNELCNT parameter information: Override Information - parameter calculation has been overridden. The calculated value was 127. The new value is 255. CHANNELCNT is not allowed to be less than 255. |