Question about DA's

I guess you already answered your own question .. DA processors have 2 ports .. 0 and 1 (if you look at them in symmwin, something I'd forget since symwin is restricted to EMC employees) or C and D if you look at them from Solution Enabler (and other tools).
In good old SCSI days SCSI processors had 4 connections in the backend (if I remember) and "C" and "D" was the active ones while "A" and "B" was the standby ones. Now Fibre DA processors have "only" 2 ports (since PBC cards do the magic) thus the name is still "C" and "D" while from an hardware point of view you have only 2 ports (0 and 1) ...

I'll ask Michael Lee to shade some light on this if possible

Hello All,

Stefano, I know you said this thread would be interesting however, it is potentially really confusing as it highlights how EMC loves to retain and re-use terminology (more on that in a moment).

Also this is not my area of expertise and someone from the EMC hardware support area could answer this with their eyes closed (of course typing this way may be more difficult ).

Anyway,as Guanabacoa has already confirmed the DA01 A0:00 and 01A:C0 are really the same thing.

As you had noted, in the earlier non-DMX Symmetrix product we had SCSI connected disk drives. These had one or 2 DA processors per backend director card. These were designated as an A and B processor. Each of these processors had a ¿C¿ & ¿D¿ SCSI bus. In fact they had a primary ¿C¿ and primary ¿D¿ and secondary ¿C¿ and secondary ¿D¿ bus ¿ these secondary SCSI connections were not configurable as they are solely for the ¿dual-initiator¿ (fault redundancy) function. I don¿t recall any references to these secondary connections as an ¿A¿ & ¿B¿ bus but otherwise your explanation is spot-on.

To summarize ¿ we have a ¿C¿ & ¿D¿ SCSI bus behind every DA processor in non-DMX models. When we introduced the DMX product range and switched to dual ported fibre channel disk drives we retained the concept of two connections behind every DA processor with up to 4 processors (i.e. A, B, C & D) per backend director card. So again you are correct we now have two fibre ports or drive loops designated ¿0¿ and ¿1¿ per DAF (DA fibre) processor.

However, for EMC terminology to remain consistent (to avoid confusion) with the older generation of SCSI based Symmetrix we retained the older ¿C¿ & ¿D¿ SCSI bus designations to keep the output from the SYMCLI consistent between families, to ensure that the format of PSE ¿Inlines¿ commands was consistent, to make the DMX SymmWin Disk Map look like a non-DMX SymmWin Disk Map (this was presumably to prevent EMC CE¿s becoming too stressed when transitioning to the DMX architecture).

So it depends where you look / what tool you use however ¿C¿ = ¿0¿ and ¿D¿ = ¿1¿ always. Finally on DMX-3 & DMX-4 the ¿C¿ and ¿D¿ SCSI bus terminology was dropped from SymmWin (to avoid confusion) as it was considered inappropriate to continue using legacy SCSI terminology (but of course the SYMCLI output didn¿t change). As long as you remember that ¿C¿ = ¿0¿ and ¿D¿ = ¿1¿ you won¿t have any issues. Finally, we have redundancy through the PBC (Port Bypass Cards) or LCC (Link Control Cards) to the dual ported fibre channel drives (the dual ports on the drives are also designated ¿A ¿& ¿B¿ but we don¿t need to go there )....

Best Regards,
Michael.

Halleleujah, I¿ve Seen the Light!!

-s-

I understand the above answer.

now we have dmx3000 and having issue with "Found DiskCage_1F:2_Right_Pbc is faulty" , needs replacement.

the affected DA s are 12a C,12b C,5a C,5b C ... EMC field eng told me that the connected 36 drives wont be available during replacement.

howcome ? we still have standby port bypass card.. right..

Hi,

       PBC do have redundancy and the disks will surely be available if the redundant PBC is working normally.

    Vipin V.K