OVMS-uid on MSA2000

[Michael Menge]

I have a MSA1000 with an embedded 2/8 fc-switch connected to an OVMS alpha server (will be replaced by an IA64 rx2660 soon).
Units are 0-11, luns are 1-12 and unit identifiers are 101-112 giving $1$dga101-112 on the vms host.
For the new MSA2000 connected to the switch I have created (via CLI>) volumes with lun 101-114 and ovms-uid 201-214 for the 1.vdisk (1.enclosure), lun 201-214 and ovms-uid 301-314 for the 2.vdisk (2.enclosure) and lun 301-314 with ovms-uid 401-414 for the 3.vdisk (3.enclosure).
Later on, using the SMU utility, even if you have selected the ovms host, you can see in the 'Manage Host Mappings' window the default mapping with the normal luns and not with the defined ovms-uids.
When creating an explicit mapping for the vms host (profile is VMS), the 'LUN*' field offers the normal lun of the volume and not the ovms-uid.
Do I have to enter the ovms-uid for all the volumes again when doing the mapping via SMU? The ovms-uid you can see only in the 'Volume Overview' window in the field 'Volume Description'. Why isn't the normal lun of the volume displayed here too?
When doing no specific mapping for the vms host (using only the default mapping), do the volumes map via the displayed normal lun or via the not displayed but defined ovms-uids?

[Shriniketan Bhagwat]

Michael,

After presenting the LUNs to OpenVMS did you execute below commands?

$ MC SYSMAN IO AUTO

This command automatically identifies and configures all hardware devices attached to the system by connecting devices and loading their drivers.

Regards,
Ketan

[Shriniketan Bhagwat]

Michael,

How is the zoning done? Does VMS server and MSA2000 are in same zone?

Regards,
Ketan

[Michael Menge]

I have disconnected the old Alpha-Server from the switch and connected the rx2660 to it. With the MSA2000 disconnected too I get on the IA64-server after $MC SYSMAN IO AUTO
devices DGA101-112 from the MSA1000 and this is o.k.
After connecting the MSA2000 to the MSA1000-switch and doing an IO AUTO again I see additional devices DGA200-214 and DGA300-314, but no devices DGA400-414 (devices with lun 100/ovms-uid 200, lun 200/ovms-uid 300 and lun 300/ovms-uid 400 I have created in the meantime with the remainig space of the virtual disks).
So where are the devices DGA400-414?
At this time only the default mapping is active. If because of that the lun numbers are used instead of the ovms-uid's, I miss devices DGA100 and DGA113-114 from the MSA2000 (DGA101-112 come from the MSA1000).
Should I do an explicit mapping of each MSA2000 device via SMU or if possible via CLI (CLI-terminal-sessions can easily be logged to save the configuration)?

[Michael Menge]

There is no zone defined, only one vms-server is connected to the switch at any time.

[Shriniketan Bhagwat]

Michael,

Below links may be helpful.

http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1303762
http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1377348


Regards,
Ketan

[Shriniketan Bhagwat]

Michael,

Please verify the steps that you followed to create the LUNs in the below link.

http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?objectID=c01984768〈=en&cc=us&taskId=&prodSeriesId=377751&prodTypeId=12169

Hope this helps to solve the problem. :-)

Regards,
Ketan

[Michael Menge]

Ketan,
to distinguish clearly between MSA1000 and MSA2000 I have deleted all MSA2000 volumes and created them new with lun=ovms-id= 200-214, 300-314 and 400-414, so unit's/unit_id's from MSA1000 and lun's/ovms_id's from MSA2000 are completely different.
When connecting the rx2660 to the switch and after $MC SYSMAN IO AUTO I see $1$dga101-112 (o.k.). After putting the cable of the MSA2000 into the switch, I see additional $1$DGA200-214 and the WWID show that these are the volumes of the 1. vdisk (1. enclusure).
The volumes of vdisk 2 and 3 are not presented. On VMS $SHOW DEV/FULL fg gives "Device FGA0:, device type KGPSA Fibre Channel" and $SHOW DEV/FULL gg shows "Device $1$GGA101:, device type COMPAQ MSA1000".

With MSU I than have made an explicit mapping of a volume VD2_V00 from the 2. vdisk (2. enclusure) to the IA64 server, but no disk $1$DGA300 is presented; same for volume VD3_V00 with lun=ovms_id=400 of the 3. vdisk (3. enclusure).
The configuration view of MSU shows as hosts the IA64-server (Discovered yes, Mapped Yes, Profile OpenVMS) and the MSA1000-switch (Discovered Yes, Mapped No, Profile OpenVMS) and an alpha-server from a previous connection (Discovered No, Mapped No, Profile OpenVMS).
So volumes DGA300-314 and 400-414 are still missing.

[Shriniketan Bhagwat]

Hi Michael,

Can you try connecting only MSA2000 to the Alpha server and try creating the LUNs on MSA2000 and check are they visible on the server without MSA1000? This does not sound like a good idea but with this, you can rule out the possibility of any conflicts between the LUNs created on both MSA1000 and MSA2000 .

Regards,
Ketan

[Michael Menge]

With suppport from HP I have created the volumes with following luns and ovms-uids:
1. vdisk VD1 lun 10-24 ovms-uid 200-214
2. vdisk VD2 lun 30-44 ovms-uid 300-314
3. vdisk VD3 lun 50-64 ovms-uid 400-414
result: all volumes could be seen by the vms host without explicit mapping.
Conclusion:
1. Explicit mapping isn't necessary if the volumes are created with appropriate luns and ovms-uids.
2. Apparently luns and ovms-uids should be different (this isn't mentioned in the docs)?
3. Can anyone explain why the volumes in the 1. enclosure (vdisk VD1) could be seen with luns = ovms-uids and the other volumes not? Are the lun numbers bounded by 2**8?