[Info-vax] Error using fibre channel target on FreeBSD from OpenVMS

[Ian Miller]

On Monday, April 25, 2022 at 11:38:16 PM UTC+1, Matthew R. Wilson wrote:
> On 2022-04-24, Robert A. Brooks <FIRST... at vmssoftware.com> wrote: 
> > On 4/24/2022 2:18 AM, Matthew R. Wilson wrote: 
> >> This could be a long shot, but I'm hoping to experiment with OpenVMS 
> >> clustering using a shared system disk over fibre channel. 
> >> 
> >> My two VMS systems are Integrity rx1600s. I'm trying to use a system 
> >> running FreeBSD as the target. After much experimentation and 
> >> troubleshooting, I've gotten pretty far, but the volumes are still 
> >> stuck "offline" in OpenVMS. 
> > 
> > I'm inferring that there is no fibre switch between VMS and the 
> > "storage array". 
> > 
> > VMS does not support FC-AL (arbitrated loop).
> There is a fibre switch (QLogic SANbox 5802V). 
> 
> It feels like it *should* be working because the OpenVMS system does see 
> the LUNs and even assigns them device names (DGA1 and DGA2 corresponding 
> to UDID 1 and UDID 2 that I assign on the target side), they're just 
> frustratingly stuck "offline" with an error. 
> 
> I've turned debug logging up pretty high on the FreeBSD side, and I'm 
> not seeing anything that looks like a SCSI command error or anything 
> like that. So I wonder if there's just some magic bit in some 
> inquire/sense data that needs to be set to make OpenVMS happy. Even if 
> there's some SCSI command that FreeBSD target support doesn't implement, 
> it'd be nice to know what command it is so that I could try to add it to 
> FreeBSD. 
> 
> Looking at various historical documentation from various SAN vendors, I 
> see many products had things like an "OpenVMS host mode". This would let 
> you set UDID, but I suspect they also know of some other values you need 
> to return in response to certain SCSI/FC commands to satisfy VMS. At 
> some point someone at SAN vendors knew what the magic was...but I can't 
> find it documented anywhere. 
> 
> Is there any way to know what the error is on the VMS side? If I run 
> ANALYZE/ERROR/ELV and then have it TRANSLATE to an output file, I don't 
> see anything obvious. I just booted the server up, and here's 
> everything: 
> 
> <https://mattwilson.org/filedrop/20220425-vms-error-log.txt> 
> 
> There are no entries with class DEVICE_ERRORS, but I do see a couple 
> TIMEOUTS and ATTENTIONS. SHOW ERROR currently reports: 
> 
> Device Error Count 
> IVMS1$DQA1: 1 
> $1$DGA1: (IVMS1 FGA) 2 
> $1$DGA2: (IVMS1 FGA) 2 
> 
> In the error log there are 4 "ATTENTIONS" and 1 "TIMEOUTS", so perhaps 
> the timeout is DQA1 and the 4 attentions correspond with the four DGA 
> errors? 
> 
> Is there a way to know what the attentions mean? "Device type 54, OS 
> flags 0x0001" doesn't really pin down the potential problem! 
> 
> Thanks, 
> Matthew
the DKLOG or FC SDA extensions on VMS may provide more clues.