What VMS systems are you managing ?

[Wim Van den Wyngaert]

The King thread made me curious about how many systems are managed by howmany people.

I would like to know :
1) what type of machine
2) how many of those machines
3) support hours 24/5, 24/7, 8/5

And how many persons are used for doing the System Management.

I have :

2 GS160=4QBB=2 clusters 24/7
4 AS4100=2 clusters 24/7
4 AS4100=2 clusters 10/5
2 AS4000=1 cluster 10/5
4 AS2100=2 clusters 10/5
1 AS2000 10/5
5 AS1000 10/5
6 AS1000 file servers 24/7
94 AStation500 10/5
and some lost stations, a bunch of decservers and the storage used by these nodes.

2 system managers doing 10 hours per day and on call 24/7.

Wim

[Kris Clippeleyr]

Wim,

Can't go in too much detail, but we manage some 80 systems (all VMS, most of them clustered (2 systems per cluster)), ranging from AlphaServer 400 to DS25, and some VAXes.
Our group consists of 3 people, working 8 hours a day. Practically, we're also on call 24/7.

Greetz,

Kris

[Willem Grooters]

Wim,

My list isn't half as big, to use an understatement):

1 PWS500au@600, VMS 7.3-2, 1-node cluster (yet), managed whenever needed but daily monitored.

System management tasks are "on request" for
- development and test machines (DS20 (1) (7.3-1), AXP1000 (2) (6.2, 7.3-1), AXP2000 (1) (7.3-1))
- a number of the 16 clusters (2100, 4100 and others, (7.1-2 - 7.3-2)) at customer's customer

and when I feel the need to do so:
- rx2600 (8.2 FT) at employers office

Not an impressive list, I know. Most experience is based on previous jobs

I'm getting curious on another issue:
If you are developing:
* What is your development system
* What languages, databases
* What development tools?

For me:

that named PWS500au (7.3-2): FORTRAN, PASCAL, COBOL, some java, some PERL, RDB and MYSQL. Using Netbeans (at least, did use standalone, working on Distributed Netbeans)

DS20 (7.3-1): SynergyDE Dibol, some java. Plain VMS (RMS, no tools but locally developed commandprocedures and conventions)

Willem

[Wim Van den Wyngaert]

Willem,

Most of the 10/5 systems are development and test systems (except 60 AStations that are production).

Developments are done in C, Sybase + stored procedures, decwindows, DSM/Mumps, a little bit of Fortran and a very little bit of Pascal. I myself use WASD and DCL and some freeware stuff.

Wim

[Ian Miller.]

mostly older small to medium VAX & Alphaserver up to some AS 4100 based clusters with HSZ70 storage at various remote sites. VMS versions from V5.2 to field test versions. My team of three does 8ish hours per day weekdays, occational weekend work and on-call 24/7. As the systems mostly belong to customers I don't think I can do details.

[Art Wiens]

1 Cluster 24/7:
4 - AS800 , 2 - VS4000-96 , 2 - VAX 4000-105

1 Cluster 24/7:
2 - ASDS10 , 1 - AS800 , 2 - VS4000-96

HSG80 SAN w/ ~ 1,224GB - 24/7

Standalones:
1 test/dev AS800 , 3 test/dev VS4000-96

~ 20 production VAX4000-105 or VS4000-96 across the country , mostly 2 node clusters - 24/7.

2 people - 24/7 ;-)

[Art Wiens]

Ooops, almost forgot:

Standalone:

1 - ES40 16/5
1 - AS800 16/5
1 - AS1200 16/5

[Bojan Nemec]

I work in a small hardware,software and consultant company (I'am one of 3 owners).

We work on several sites with different systems from VAX 3100 to Alpha DS25, some of them are in cluster with different storage subsistems (from ordinary SCSI discs to HSV storage). Operating system version also varies from 5.5-2H4 to 7.3-2 in total about 20 or more systems.

On most of them, we are doing System Management and writing application software (partial solutions or full IT support). Some of the sites are 24/7 per call.

For this (system management and application software) we are in 2. For hardware and networking there is a third guy, but we use to mix our knowledge.

Bojan

[Jan van den Ende]

Hi,

our environment is a little bit different again:
ONE production cluster, and 1, somtimes 2 test systems.
Productioncluster was build as 2 AS2100.
Add a VaxStation 4000-90. Add 3rd AS2100.
Move 2 AS2100 to second site at 7KM to form multi-site cluster.
Add ES40, Upgrade AS2100's to AS2100A's. Remove VaxStation. Add ES45 Remover AS2100A Add ES40. Remove AS2100A. Add ES45. Remove AS2100A. Add DS15Station. and lately: change DS15 station into server, and remove ES45
(ES45 "temporarily" available for prestige project).
All this while VMS 6.2 -> 7.1-2 -> 7.2-1 -> 7.3-1 -> 7.3-2.
Running DBMS, Rdb, Sybase, Progress, RMS, Basis+, and "random access"-(ported from Prime) Databases.
24 * 365.25, available, 7 * 24 support.
Support responsetime for two applics "should be less than 15 minutes".
6500 authorized users.
200 - 300 weekly mutations (major or minor) to user base, about same number in application authorisation changes.


- anybody seconding the plan to tranfer this to a Tru64 - M$ configuration?

Support on VMS and on the technical sides of all applics com from 3 people, one of them 50 % of time.

Jan

[Zahid Ghani]

I have 4 node Disaster Tolerant Cluster
2 x nodes (DS25) at each site
HSG80 Storage
with LES100 intersite link
Robomon cluster management.
and MSL5026 tape library.
24x7 support provided by 3member team.
with average of 1500 users connected to the cluster at any moment.
Application is SOP, WEB ordering,logistics- all bespoke.
zahid

[labadie_1]

Jan


Support responsetime for two applics "should be less than 15 minutes".
6500 authorized users.
200 - 300 weekly mutations (major or minor) to user base, about same number in application authorisation changes.



Support on VMS and on the technical sides of all applics com from 3 people, one of them 50 % of time.


So two and a half guys to manage all that ?
Congratulations.

I suspect that after the move to Tru64 or Windaube, it will require a lot more people.

I wonder how CFO and head of IT can justify that they need a lot more money to do the same thing (not counting the migration cost).

Gerard, puzzled :-(, but having already seen very similar things several times

[Karl Rohwedder]

I am freelancing together with 3 friends and I am the last VMS guy here, all others are doing
mainly Windows...
On the customer site I am managing a bunch of 10 VAX systems, mainly VAXstations and MicroVAX and 2 VAX4000-400 and 14 Alphas, mainly
DS10/DS20 and some older ASRV 1000A system, some clustered some standalone.

[Wim Van den Wyngaert]

I forgot to mention : 1 person is 60%, I also manage some smelly middleware and monitor my systems 10/5. Only backups and monitoring at night is not our business (repairing is). Of course I don't count all my bosses but 20% of our work is for satisfying them instead of VMS.

Jan : we have a menudriver to maintain users. It takes 1 minute to do the work and 1 minute for admin. So 200 users would take 400 minutes or about 7 hours.

I am amazed about the differences in load per person ...

Wim

[Steven Underwood]

Well, I'm am no where as large as many of you here but...

1 ES40 24/7 (moving to 7.3-2)
1 DEC3000/400S 24/7 (7.1)
1 MicroVAX II 8/5 (5.2)

Just me here to maintain these plus all the other equipment on our network (~300 PC's). So glad VMS does not need much babysitting ;)

[Jan van den Ende]

Wim,

basically (nearly) all of usermanagement changes are performed by a batchjob that gets its info from the mutations in the personnel system.
Those just require us < 5 minutes / day to check smooth running.
We have convinced Management that ANY info in the personnel system is considered THE info our systems have to go by, so ANY errors (in spelling of name, in end-date of job/contract, in department...) that are found in error ONLY get repaired in the personnel system, and next morning early it is fed into VMS.
Btw: different applications DO need info from the personnel system that are NOT in SYSUAF, and that all is maintained by the same job (5 - 15 minutes daily).

Changes in authorisation for (level of) application use are REQUESTED by a user's superior, and decentrally there exist several dozen Authorisation Officers, who have to sign the request. Then it is transfered to a department that inplements them (over VMS, Tru64, AIX, Novell & W2K)
In VMS application authorisation is implemented as an identifier _ACCESS, which governs the appearance of the applic in the main menu, and the protection of the applic funtion files. Granting/revoking this identifier is coupled (by Advanced Server) to adding/removing corresponding group membership in Novell & W2K, which then governs the availability of the app. on the desktop.
The people authorized to do those autorisations are authorized to access the AUTOR applic, and those users have NO privs, the AUTOR applic has the necessary ones.
Of course all this with nice accountings and reportings.
So, it WAS quite a job to put this all together (it has been a years-long expanding development), but NOW it requires very little effort.

Although: a big reorganisation, and a change of personnel system, are to be implemented before year's end, so, never a dull moment.

Cheers,
have one on me!

Jan

[Lawrence Czlapinski]

Wim,
2 sites
4 clusters (2 production and 2 development)
10 Standalone VMS (8 production and 2 development)
1 Unix (Alpha or Vax)
2 system managers doing up to average of 8 hours per weekday plus on call 24/7

[Martin P.J. Zinser]

Hi Wim,

while I can not go into details, about a dozen production clusters, a few hundred VMS systems in other use, SAN, PC and Unix server support.
About 15 employees in system management.

Greetings, Martin

P.S. My main affiliation is actually not with that part of our operations ;-)

[Willem Grooters]

Jan forgot to mention one thing: The cluster that is up-and-running-and-available for over 7 years now, is part of that environment. Without interruption (of some significance). ALL hardware and software replacement were done 'on the fly'.

Only he can tell if that involved more people than the two-and-a-half. But I doubt it would.

Who can _guarantee_ this availability with Unix or Windows - and the TCO....

Willem

[labadie_1]

Wim

may be we should start un new thread about Vms TCO... :-)

[Ian Miller.]

The VMS advantage is often hidden as the true cost of supporting systems running inferior operating systems is not known (too much unpaid overtime by the Slaves of the Dark One). The reliability adjusted TCO (as seen in http://h71000.www7.hp.com/openvms/whitepapers/tco_clusters/TCO_WP_Feb04.pdf ) is an interesting approach with I think shows the true cost more accurately.

I think the figures shown here (people/server) could only be matched by a very homogenuous and tightly controlled unix server farm with a lot of automation.

[Jan van den Ende]

To complete Willem's extra:

Yes we planned and executed it with our own team.
I guess none of us would dare let someone else in on such things, before we had assured ourselves thoroughly that that person DID fully understand our configuration, and the somewhat more sensitive points therein.

What I also still forgot in our migration path: just before the first ES45 entered, a dual-site double-redundant HSG80 SAN was added, only attached to ES'ses, and on the fly ALL data was gradually moved there, from the old HSZ40-attached disks.
When all AS2100A's where removed (their system disks the last disks used on HSZ disks), the HSZ40's and those disks were removed.

Jan

[Antoniov.]

Wim,
I'm owner of a very little software house. I manage directly:
1 AS400 V6.2
1 AS400 V7.3-2
1 VAX3100 V5.2 (not yet used)
On october I'll have
1 RX2600 V8.2
I manage my customers systems:
Some DS10 V7.1-2 ... V7.2-1
Some AS400 V6.2
Some AS800 V7.1-2
Some AS1200 V7.1-2 ... V7.2-1
No cluster.

Antonio Vigliotti

[John H. Reinhardt]

Hi all,

At the moment I don't manage any VMS systems professionally, however, I do attempt to keep myself active with my home systems:

2) AS800's clustered (V7.3-2)
2) AS1200's clustered (V7.3-2)
1) PWS500au (V7.3, V7.3-2, others as required)

Storagewise:

2) HSG80 w 436MB (RA8000)
2) HSG80 w 464MB (M2200/3x4254)

[Dieter Rossbach]

alphaserver:

cluster: 2x as1200 /hsz50 / 24x7

cluster: DS25/DS20E/MSA1000 / 12x5

cluster as800/as1000/as1200 / 12x5

cluster 2x as4100 /hsz50 / 24x6

2 x as2100 / 24x6

ds10
ws XP1000
ws 500au
ws 600au
4 x ws200

vax:

2x 4000 / 24x6
6x 3100 / 24x6
1x SIMH

mostly VMS with:

5.5-2, 6.1, 7.1, 7.3 ,7.3-1, 7.3-2

RdB 7, Oracle 7,8,9 Ingres II 2.0
Apache
Tomcat
JAVA

2 Debian systems with exim4, spamassassin, clamav, cvs

same network stuff (Hub900, Cisco- and bintec-Routers), some PCs

1 System manager (me) and a backup, just in case ...


Dieter

[Art Wiens]

I've managed to get over the fact that we still have to use VaxStations w/ VMS v5.5-2 and v6.2, but to Steve Underwood - a MicroVAX II ?!?! Does it still have RD53's, a TK50 drive and a LA120 console?

Is that a bet with someone that it will/won't run forever?!?! ;-)

You should get a hat for that!

Cheers,
Art