| Title: | + OpenVMS Clusters - The best clusters in the world! + |
| Notice: | This conference is COMPANY CONFIDENTIAL. See #1.3 |
| Moderator: | PROXY::MOORE |
| Created: | Fri Aug 26 1988 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 5320 |
| Total number of notes: | 23384 |
Hello!
We have this customer who has a AS8400 and AS 4100. The AS8400 is
running OpenVMS V6.2 while the AS4100 is running V6.2-1H3. The AS8400
is connected to a starcoupler with an HSJ40 on SW 800 while the AS4100
has a KZPSC-BA controller connected to 2 storage bays. Both of them are
connected to the FDDI ring.
Question:
1. How do we configure this set up as cluster thru FDDI when there is
no common disk to set as quorum disk? I set the votes of AS8400 as 2
and 1 for AS4100 with exepcted votes=2; is this valid?
2. Will this two communicate - mixed versions?
Thanks.
Noel gesmundo
Digital/Philippines
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 5282.1 | Two Node VMScluster Config, No Shared Storage... | XDELTA::HOFFMAN | Steve, OpenVMS Engineering | Mon Apr 14 1997 11:44 | 39 |
: 1. How do we configure this set up as cluster thru FDDI when there is
: no common disk to set as quorum disk? I set the votes of AS8400 as 2
: and 1 for AS4100 with exepcted votes=2; is this valid?
If there is no common disk accessable to both systems, then there
is no point in using a quorum disk.
I'd recommend taking a look at the documentation around the quorum
scheme -- in this case, a two node VMScluster with no multi-path
or multi-port shared storage, the only possible configurations are:
o primary and secondary -- only one node has votes, the
non-voting node can boot and fail whenever...
o peer configuration -- where the failure of either node
will cause a quorum hang on other.
o non-VMScluster -- where the two nodes are not configured
in a VMScluster.
If the customer were to upgrade to V6.2-1H3 or (better) V7.1, the
customer could configure a shared SCSI bus, assuming the correct
SCSI hardware is present and can be configured. (V7.1 adds support
for port allocation classes, which makes shared SCSI far easier.)
Alternatively, the customer could configure a CI adapter on the
AlphaServer 4100.
: 2. Will this two communicate - mixed versions?
Yes. When asking these questions, please see the Software Product
Description -- the SPD. The SPD contains the matrix of versions
that are supported for mixed-version operations for "migration"
and for "warrantee" support.
--
Also consider issuing a NOTES> command such as...
MODIFY NOTE 5282.0/TITLE="Cluster Quorum Config Questions"
| |||||
| 5282.2 | STAR::NCARR | Talk dates & features - but never together.... | Mon Apr 14 1997 14:39 | 6 | |
Clearly the best thing to do here is put a CIPCA in the 4100 and connect it to the star coupler. Anything else is going to be much more messy and less useful. Note that there are now two versions of the CIPCA - one that uses two PCI slots, and one that uses one PCI slot + one EISA slot. | |||||
| 5282.3 | systems clustered with different votes. | MANM01::NOELGESMUNDO | Wed Apr 16 1997 07:23 | 17 | |
Hi!
Thank you for your replies. We configured the systems as clustered by
setting 2 votes for the AS8400 and 1 vote for AS4100 with expected
votes=3. This is a temporary solution since they have already bought a
CI adapter for the AS4100.
Obviously for this configuration, the AS4100 can't boot as a one-node
cluster but the AS8400 can. The only feature that is important now is
that disks can be shared across the 2 systems. I already explained this
to the customer and agreed to expedite the delivery of the CI adapter
for the AS4100.
Thanks.
Noel
| |||||