I got some issues with the LSAN environment, where I have the tapedrives been accessed by the media server across the site. We have dual fabric and right now, we have issues with one fabric where , the media server is not been able to access any drives across another DC. But the other fabric is working fine without any issues.
I have already checked the config and compared them , and they look exactly same . Also the physical devices are imported and exported well.
Another test we did , was to cable another test server and allocated the disk over this LSAN and it again sees the paths only through one fabric. So maybe this issue is within DWDM/Switch Config.
Could you please give me some more ideas , where to look for the issues.
I think you already hit on the number one place I would look... that was to make sure that all the devices you expected show up as being exported properly into the edge fabric where it is being accessed. If that is all working properly and you can see the exported devices then there may be a deeper issue that will require a Service Request to dig into.
I've only ever had FC Routing problems where the devices don't get exported properly. Once they appear in the other edge fabric connectivity is pretty much a given (unless there is an incompatibility between the target and initiator configuration - i.e. it wouldn't work if they were connected to the same fabric anyway)
pretty weird situation we have...the fabric which had the problem is connected via two isl ports to the adva components and if both ports are up, we can't see the tape drives , but if one port is enabled or up ( any of the two ports ), the tape drives are visible. We had trunking and we disabled that and , to me it seems some dwdm config issue...we are using 5TCE cards which is recommended to run on transparent mode , we had it on 4g framed mode + with diff firmware on two sides of the same fabric, so hopefully after changing this to proper setup..we might find a solution.
Have you doublechecked to make sure that both of the EX ports for remote fabric connectivity are configured with exactly the same edge fabric ID? I've never tried to connect to a single fabric with two different FIDs so I don't know what would happen, but I know it would be a problem...
Yes, we did the checks , also had checked it by vendor engineering...but seems now it is pointing to a routing issue internally within the switch internal logic...to be honest , I don't have a full view of what it means as yet since engineering in brocade took over the case..I will put information for all when I have it..
Thanks for you time on this...appreciate it.
Having it go to engineering is one of those things that is a good thing and a bad thing. On the one hand you know if it gets there you probably didn't make any stupid mistakes in the configuration (and if you did then other people missed it too) 🙂 That's always a nice thing to know.
On the other hand, if it gets to engineering there is no telling how long it will take to track down and what kind of fix will be required. Our very first FC Routing implementation failed due to a bug in the FOS code that normally wouldn't have been a problem except for the specific configuration (switch domain IDs and edge fabric IDs) that Iselected. Of course... since it was my very first attempt at FCR my assumption was that I was doing something wrong.
I look forward to hearing the results. Every little bit of extra knowledge we can get on these systems helps in the future!