Booting without joining a cluster?

[Phillip Tusa]

Greetings to all ...

We have 2 VAX 4000-200 system clustered together with VMS 6.2. We are trying to
boot one of the nodes WITHOUT joining a VAX cluster. Is this possible to do this without booting stand-alone? If so, could someone post the >>>B ??? command?

Thanks in advance!

Phil

[Robert Gezelter · Accepted solution]

Phil,

You will have to use a dedicated system disk. No cluster means NO SHARED devices.

The general command is BOOT or B . I do not recall offhand which is used in the VAX 4000-200 firmware, and I do not have the manual handy at this instant.

You will also need a non-clustered system root, or you will have to boot conversationally to disable clustering.

I am in the office today, if you want to speak on the phone.

- Bob Gezelter, http://www.rlgsc.com

[Phillip Tusa]

>You will also need a non-clustered system >root, or you will have to boot >conversationally to disable clustering.

Bob, Thanks for your quick reply! Is there a HP document online that would describe the "conversational boot"? I've read about this before but it's been a while for me.

Thanks!

Phil

[Jim_McKinney]

BOOT /R5:1 ddcu

will initiate a conversational bootstrap and give you a SYSBOOT> prompt - from there you can alter any GEN parameters that you desire priot to CONTINUEing. Heed Bob's warning about sharing devices - don't let it happen if your system is not clustered.

[Steven Schweda]

> [...] I've read about this before [...]

Probably because it's a frequently asked
question. Wait! Perhaps it's in the FAQ:

http://www.hoffmanlabs.com/vmsfaq/vmsfaq.txt

A Google search for something obvious, like:
vms OR openvms conversational boot
would lead you to other useful bits, too,
such as a VMS Upgrade and Installation
Manual (or several). Throw in VAX to limit
it more.

[Jan van den Ende]

Phil,

this _IS_ a bit dependent on your hardware config.

_IF_ you have two system disks,
_THEN_ Jim's answer is your answer. his ddcu is your local system disk.

_IF_ you have a single (common) system disk, then your _ONLY VALID_ option is to use the installation CD (I don't remember if that does exist for Vax at 6.2? Otherwise, do you have a higher version; that will work as well). The ddcu in Jim's command then is your CD drive,.

Success!

Proost.

Have one on me.

jpe

[Jan van den Ende]

Philip,

from your Forum Profile:


I have assigned points to 18 of 54 responses to my questions.

Those do date back almost one year.

Maybe you can find some time to do some assigning?

http://forums1.itrc.hp.com/service/forums/helptips.do?#33

Mind, I do NOT say you necessarily need to give lots of points. It is fully up to _YOU_ to decide how many. If you consider an answer is not deserving any points, you can also assign 0 ( = zero ) points, and then that answer will no longer be counted as unassigned.
Consider, that every poster took at least the trouble of posting for you!

To easily find your streams with unassigned points, click your own name somewhere.
This will bring up your profile.
Near the bottom of that page, under the caption "My Question(s)" you will find "questions or topics with unassigned points " Clicking that will give all, and only, your questions that still have unassigned postings.
If you have closed some of those streams, you must "Reopen" them to "Submit points". (After which you can "Close" again)

Do not forget to explicitly activate "Submit points", or your effort gets lost again!!

Thanks on behalf of your Forum colleagues.

PS. - nothing personal in this. I try to post it to everyone with this kind of assignment ratio in this forum. If you have received a posting like this before - please do not take offence - none is intended!

PPS. - Zero points for this.

Proost.

Have one on me.

jpe

[Robert Gezelter]

Phil,

Perhaps a clarification would be appropriate.

What is the particular reason that the machine needs to be booted WITHOUT being a membeer of cluster?

There are often more options in these situations than are apparent at first examination.

- Bob Gezelter, http://www.rlgsc.com

[Hoff]

If there are shared storage or shared cluster communications interconnects here, then bad things can happen. ("Bad" here can involve system crashes on one or both nodes, and can potentially involve data corruptions within storage.)

As for conversational bootstraps and related topics, yep. That sort of detail is in the OpenVMS manuals and in the OpenVMS FAQ. Alternatively, there are topics posted on bootstraps and clustering and otherwise at sites including the following:

http://64.223.189.234/node/204
http://64.223.189.234/node/192
http://64.223.189.234/node/169

Google is your friend, too.

Stephen Hoffman
HoffmanLabs LLC

[Jon Pinkley]

Phillip,

What you are planning to do is risky if you have any storage that is directly connected to both VAX 4000-200 systems. I.e. if you have disk shared on a DSSI or some other shared storage (CI, shared SCSI etc.)

If the only way the disk are shared is via MSCP over ethernet, then it is less risky, but not recommended.

Clustering provides the co-ordination of sharing the disk. If one node boots without being in the cluster, it will think it has complete control of any disks it is able to see. The remaining node in the cluster (if there is a quorum disk) will also think it is the only thing accessing the disk. This quickly leads to disk corruption, and cleaning up the mess is a lot harder than avoiding it in the first place.

So back to Bob's question: What problem are you trying to solve by booting a node while it isn't a part of the cluster?

Do the two VAX 4000-200s currently boot off the same system disk? If so, as Jan hinted, a conversational boot to turning off clustering is almost a guarantee of a trashed system disk, if the other node is currently running.

So you will need to boot off some other system disk, and avoid mounting any of the shared devices. If you have a local drive with a copy of your current system device, that could be used, but you would need to make sure that it does not mount any disks that are on shared storage interconnects.

Jon

[Jan van den Ende]

Phil,

please _DO_ give due attention to Hoff's and Jon's warnings.
OTOH, this exact path is what allows for, and is fully supported for, the so-called "rolling upgrade".
- (if common system disk:) get a current COPY of the system disk.
- boot one system off of that WITHOUT joining the cluster AND WITHOUT MOUNTING ANY OTHER DISKS!
- on this system do the upgrade.
- boot into the cluster off that upgraded disk (you now have a mixed-version cluster)
- if common disk, (re-)boot other system(s) off of the upgraded disk; if separate system disks, repeat for each SD.

So YES, the concept is well known, BUT, strict precautions are necessary, and the "standalone" system should be used for NOTHING else!

hth

Proost.

Have one on me.

jpe

[Robert Gezelter]

Jan, Jon, Hoff, et al.,

I have had a sidebar conversation with Phil, and uncovered some details concerning his situation that did not quite come out in the posting that started this thread.

Apparently, the situation involves a hardware failure involving the second CPU, and a cluster that was reduced from three to two nodes without any changes in the voting parameters.

I talked him through the procedures for a conversational boot, and strongly cautioned concerning bringing the second system back, if it is repaired.

- Bob Gezelter, http://www.rlgsc.com

[John Gillings]

>(I don't remember if that does exist for
>Vax at 6.2?

The same Standalone DCL available for Alpha (the $$$ prompt) is also present on VAX distribution CDs, but curiously not well documented. I'm fairly sure it goes back as far as V6.2.

Boot from root 1 on a CD drive or image backup of a distribution CD on another disk:

>>> B/R5:10000000 DKA400

Even if the V6.2 CD doesn't have $$$, there's no reason you can't boot a later version. Just avoid doing anything that isn't back compatible to V6.2 (for example initializing disks with recent features like small cluster sizes on large disks enabled)

[Mrityunjoy Kundu]

perform a conversational boot...
>>>boot/R5:0,1 DUA0:
you will get sysboot prompt
sysboot>set VAXCluster 0
sysboot>continue

[Hoff]

There are some considerations around that previous reply here pointing to VAXCLUSTER parameter and the conversational bootstrap.

-- the cited approach appears incomplete for its intended purpose (you will also want to set NISCS_LOAD_PEA0 to 0, as various versions of OpenVMS can treat these parameters as a pair, and as you'll probably want to cancel the parameter write); failure to have at least NISCS_LOAD_PEA0 can leave you in an odd state.

-- the approach of disabling the cluster when shared interconnections around can be an exceedingly hazardous answer (for the reasons cited previously in this thread; I've corrupted disks during testing, and have seen crashes due to driver confusion),

-- and the VAX console boot command syntax shown seems odd and seems to be mixing Alpha SRM and VAX syntax; AFAIK, VAX expects "BOOT /R5:10000001 ddcu" for a VAX console conversational bootstrap.)

The OpenVMS FAQ has sequences which include references to various of the console and platform permutations, as well. This in addition to convenience features, such as skipping the parameter write operation.

But -- as Bob G. mentioned a couple of posts back -- this bootstrap issue has been resolved off-line.