- Community Home
- >
- Servers and Operating Systems
- >
- Operating Systems
- >
- Operating System - OpenVMS
- >
- Re: 3 Site VMS Cluster with Shadowing and Quorum S...
Categories
Company
Local Language
Forums
Discussions
Forums
- Data Protection and Retention
- Entry Storage Systems
- Legacy
- Midrange and Enterprise Storage
- Storage Networking
- HPE Nimble Storage
Discussions
Discussions
Discussions
Forums
Forums
Discussions
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
- BladeSystem Infrastructure and Application Solutions
- Appliance Servers
- Alpha Servers
- BackOffice Products
- Internet Products
- HPE 9000 and HPE e3000 Servers
- Networking
- Netservers
- Secure OS Software for Linux
- Server Management (Insight Manager 7)
- Windows Server 2003
- Operating System - Tru64 Unix
- ProLiant Deployment and Provisioning
- Linux-Based Community / Regional
- Microsoft System Center Integration
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Community
Resources
Forums
Blogs
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО03-23-2009 12:30 PM
тАО03-23-2009 12:30 PM
Question regarding the configuration of a Quorum Server on in a 3 site Cluster using Shadowing.
We project to use :
- 1 ES47, 1 ES80 and 2 Blade BL870c on site 1 ;
- 1 EVA8100 on site 1 ;
- 1 ES47, 1 ES80 and 2 Blade BL870c on site 2 ;
- 1 EVA8100 on site 2 ;
- 2 FC Fabric with 2 x 4Gb/s links each ;
- 1 private NSPOF LAN for SCS, DecNEt ;
- Intersite LAN link 10 Gb/s ;
- On site 1 and 2 there is 2 systems disk;
- 1 for Alpha ;
- 1 for Itanium ;
- SYSUAF is on a common disk ;
- 1 BL860c use as a Quorum Server on site 3 ;
- This server have is own local disks ;
Is it better :
1) To configure the Quorum Server just using is own local disk and sharing the SCS and DecNet communication between the others nodes of the Cluster ?
2) Then, to configure the Quorum Server with is own SYSUAF file instead of mounting the common disk of the cluster ?
3) If we use a separate SYSUAF file on Quorum Server, do the SYSMAN command "SET ENVIRONMENT/CLUSTER" will prompt us for the password of the "system" account ?
3.1) If we use a separate SYSUAF file, since SYSMAN use DecNET, setting a PROXY account in SYSUAF file on the Cluster side and on the Quorum Server side be a way to resolve the prompt of a password ?
YvEs H.
Solved! Go to Solution.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО03-23-2009 12:52 PM
тАО03-23-2009 12:52 PM
Re: 3 Site VMS Cluster with Shadowing and Quorum Site
If you're going to go to the trouble of having a quorum site, you should go the whole way!
I'd expect the quorum node to have its own, local system disk. Maybe locally mirrored (controller based?). The goal is to make your system disk effectively read only.
ALL your cluster environment files (there are LOTS more files than just SYSUAF, see the comments in SYLOGICALS.TEMPLATE) should be on a common disk shared across the cluster. The ideal would be a three member shadow set, one member physically on each site. For the paranoid, each shadow member would also be controller mirrored. The storage doesn't need to be big, or particularly fast.
This set of files defines your cluster. In theory, you should be able to recover your whole system by clean installing OpenVMS from distribution media, then pointing the logical names at your environment files (the procedure for doing this should be on the common disk).
Make sure you have a backup for your cluster data. I like to keep a copy of every file on the same disk, then a second set of copies on another disk, from which they can be sent to tape.
Having separate files will just cause confusion and make any recovery procedures more complex and error prone. Recoveries are stressful enough without having the added complication of trying to resolve mismatched, expired or forgotten passwords.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО03-23-2009 01:50 PM
тАО03-23-2009 01:50 PM
Re: 3 Site VMS Cluster with Shadowing and Quorum Site
If you do decide to run distinct and parallel SYSUAF files (rather than shadow-level copies), you need to coordinate UIC and identifier values across all hosts. Which means any usernames that occur in more than one SYSUAF file must have the same UIC, and the same for matching identifiers in the RIGHTSLIST.
I'd consider an limited SYSUAF file here via the SYSUAFALT file and the UAFALTERNATE parameter (which is where I think you're going), but that's irrespective of John's comments. This SYSUAFALT file here serves strictly as a backstop; I tend to use that as a fast way to reboot with only limited server access among the user community. As John comments, I'd also run with the same shared SYSUAF file in production.
And if you're asking these (good) questions, the next series of questions involves the "what have you missed?" discussion in your DT and archival and system configuration and policies and procedures and administrative processes.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО03-23-2009 01:50 PM
тАО03-23-2009 01:50 PM
Re: 3 Site VMS Cluster with Shadowing and Quorum Site
If the quorum disk is locally attached,
it serves no purpose. It simply adds votes
if that system is up and adds to the load.
A shadow disk can not be used as a quorum disk, (however you can use mirrored disks that appear to VMS as one volume.
The disk should able to be seen on all systems and not shadowed, and the allocation class for it must not be zero.
Bob Comarow
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО03-23-2009 02:16 PM
тАО03-23-2009 02:16 PM
Re: 3 Site VMS Cluster with Shadowing and Quorum Site
DO heed Hoff's remark about "what have we missed"
Keith Parris has a lot of stuff on that, and it is much easier to follow existing tracks than trying to find ALL yourself (but if you are planning this stuff, you probably already got there.
One more thing on redundant shadow members:
John wrote
>>>
For the paranoid, each shadow member would also be controller mirrored.
<<<
It may be worth mentioning that recently the 3-member limit was relaxed, you can now have 2 members at each site (even with facility for an extra member to split off for backup purposes)
hth
Proost.
Have one on me.
jpe
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО03-23-2009 02:28 PM
тАО03-23-2009 02:28 PM
Re: 3 Site VMS Cluster with Shadowing and Quorum Site
>It may be worth mentioning that recently
>the 3-member limit was relaxed, you can
>now have 2 members at each site
News to me! I knew this was planned, but didn't notice it actually get released. What's the minimum version & patch level that supports >3 members?
The help file on V8.3-1H1 still says "Binds up to three physical devices into a shadow set..."
re: Bob,
There is no quorum disk here, it's a quorum node (or, indeed a quorum site).
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО03-23-2009 06:20 PM
тАО03-23-2009 06:20 PM
SolutionMy site sounds cunningly similar to what you want to achieve. 2 x rx2620 as the production nodes. 1x Alpha DS10 as quorum node. All in different machine rooms in different buildings around the site. rx2620s create shadow sets of one or more disks on each server. DS10 does not mount any disk that is mounted on the rx2620s.
DS10 has its own sysuaf and rightslist, partly because I do not want users logging in to the the ds10. (Yes, I know I could have smarts built into the sylogin, but why bother?)
I have proxys between the ds10 and rx2620s. SYSMAN works happily between all three nodes without prompting for a password.
Seems stable enough, cluster was formed Oct 12 2005 and hasn't been down since.
Have fun.
PJ
Peejay
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
If it can't be done with a VT220, who needs it?
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО03-24-2009 05:13 AM
тАО03-24-2009 05:13 AM
Re: 3 Site VMS Cluster with Shadowing and Quorum Site
Shadowing with > 3 members will work in V8.4:
$ mou/sys dsa0:/shad=($1$dkb0:,$1$dkb100:,$1$dkb200:,$1$dkb300:) s1
%MOUNT-I-MOUNTED, S1 mounted on _DSA0:
%MOUNT-I-SHDWMEMSUCC, _$1$DKB0: (THEALP) is now a valid member of the shadow set
%MOUNT-I-SHDWMEMCOPY, _$1$DKB100: (THEALP) added to the shadow set with a copy operation
%MOUNT-I-SHDWMEMCOPY, _$1$DKB200: (THEALP) added to the shadow set with a copy operation
%MOUNT-I-SHDWMEMCOPY, _$1$DKB300: (THEALP) added to the shadow set with a copy operation
%%%%%%%%%%% OPCOM 24-MAR-2009 11:08:51.82 %%%%%%%%%%%
Message from user SYSTEM on THEALP
%SHADOW_SERVER-I-SSRVBEGCPY, BEGINNING copy operation on _DSA0: at LBN: 0, I/O size: 127 blocks, ID number: 12000227
$ sh dev dsa0
Device Device Error Volume Free Trans Mnt
Name Status Count Label Blocks Count Cnt
DSA0: Mounted 0 S1 514448 1 1
$1$DKB0: (THEALP) ShadowSetMember 0 (member of DSA0:)
$1$DKB100: (THEALP) ShadowCopying 0 (copy trgt DSA0: 15% copied)
$1$DKB200: (THEALP) ShadowCopying 0 (copy trgt DSA0: 15% copied)
$1$DKB300: (THEALP) ShadowCopying 0 (copy trgt DSA0: 15% copied)
Fwiw,
Jur.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО03-24-2009 12:34 PM
тАО03-24-2009 12:34 PM
Re: 3 Site VMS Cluster with Shadowing and Quorum Site
YH