| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 2049.1 | | KITCHE::schott | Eric R. Schott USG Product Management | Wed May 07 1997 16:54 | 6 |
| Hi
It is not supported...can you tell us what they are trying to achieve
or what problem they want to solve?
|
| 2049.2 | Customer mistrust on TruCluster solution | ROMTSS::MORRONE | | Thu May 08 1997 06:17 | 24 |
|
Eric,
Thanks for your prompt reply.
Even it is not expressly stated within any docs, I deduced it's not supported;
in fact, I never suggested such configuration to any Customer.
The Customer, requesting that unusual configuration, would like to have the
chance to access HSZ40 disk drives managed by TruCluster A from TruCluster B
in the case TruCluster A is not anymore able to provide access to its storage;
so that, at the time of the problem, instead of physically connect a Diff.
SCSI cable to manage TruCluster A's storage from TruCluster B Customer will
have a SCSI connection ready to be used by the surviving TruCluster.
At the base of this strange request, from the Customer point of view there is
a lack of trust toward a single '4 Turbolaser members configuration' (..based
on bad Customer experiences on DECsafe/ASE environment reliability..).
Could you please confirm the unsupported configuration is not valid within a
DECsafe/ASE environment as well.
Thanks indeed again.
Regards, Maurizio
|
| 2049.3 | | KITCHE::schott | Eric R. Schott USG Product Management | Thu May 08 1997 13:49 | 25 |
| >
>Even it is not expressly stated within any docs, I deduced it's not supported;
>in fact, I never suggested such configuration to any Customer.
>
>The Customer, requesting that unusual configuration, would like to have the
>chance to access HSZ40 disk drives managed by TruCluster A from TruCluster B
>in the case TruCluster A is not anymore able to provide access to its storage;
>so that, at the time of the problem, instead of physically connect a Diff.
>SCSI cable to manage TruCluster A's storage from TruCluster B Customer will
>have a SCSI connection ready to be used by the surviving TruCluster.
>At the base of this strange request, from the Customer point of view there is
>a lack of trust toward a single '4 Turbolaser members configuration' (..based
>on bad Customer experiences on DECsafe/ASE environment reliability..).
>
>Could you please confirm the unsupported configuration is not valid within a
>DECsafe/ASE environment as well.
It is not supported...We have not tested it and have no idea
how it might work.
Why if they are willing to have all the nodes on the same shared
buses would they not put them in the same cluster??
>
|
| 2049.4 | Customer doesn't trust yet any 'Cluster solution' | ROMTSS::MORRONE | | Thu May 08 1997 17:48 | 14 |
|
Eric,
As I tried to tell you, at the moment Customer doesn't completely trust
a single 4 members TruCluster solution; he thinks there are chances that
a SW or HW problem within any 'TruCluster component' would make inoperative
all 4 systems (instead of two as in the previous configuration).
Anyway, it's at least 2 months I'm trying to convince Customer to implement
the 4 members TruCluster solution as the best configuration for their needs.
I'll let you know which will be the end of the story...
Thanks a lot again for your cooperation.
Regards, Maurizio
|
| 2049.5 | ??? | BACHUS::DEVOS | Manu Devos NSIS Brussels 856-7539 | Fri May 09 1997 16:57 | 20 |
| Maurizio,
If the customer is not convinced by the robustness of ASE, maybe you
can try to explain that even if ASE is not able to start or fail a
service over, you can mount it manually with a script that I made public
in this notefile: "mount_service". Do a dir /tile="mount".
ASE is not magic. It simply place some storage online, then start this
storage (LSM) and finally mount the storage on mount-point. Thus, if
ASE is down, you can still use the storage manually.
Maybe, your customer can accept this 4 nodes config if he knows that
his data can be accessed without ASE.
The only precaution is to have terminators on Y-cables at the ends of
the bus to allow the repair of any component of this bus and of course
to have your data mirrored across two different buses. That means : no
single point of failure.
Manu.
|