StoreEver Tape Storage
1748208 Members
2804 Online
108759 Solutions
New Discussion юеВ

NSR e1200-320 4G and AIX fabric zoning

 
Joe Profaizer
Super Advisor

NSR e1200-320 4G and AIX fabric zoning

We seem to have a common denominator in that most if not all our AIX hosts are having issues with our MSL6060. The MSL6060 has four LTO4 tape drives (two each controlled by separate NSRs). All the zoning is done correctly and I can see the WWNs from the mapping menu. By the way, we have no issues with HP-UX hosts nor Windows hosts. ;-)

I practice using 'Port 0 Device Map' for all our tape drive SAN hosting and it seems to work well. It's just that AIX sees the drive for a bit, and it may work for a while, then the AIX host's device will change from an 'available' state (which is good); to a 'defined' state (which is not good as it eliminates capability to use the drive).

Is there any special settings on the NSRs that we should use for AIX zoning only? It just seems that AIX is the common denominator here. Again, all HP-UX and Window's tape drives are accessible via the SAN.

Thanks in advance.

..Joe
4 REPLIES 4
Marino Meloni_1
Honored Contributor

Re: NSR e1200-320 4G and AIX fabric zoning

I'm not completely sure about your config, but:
-if you have one NSR per LTO drive
>have a look at the nsr log and see if you have unexpected reboot logged. Having two NSR per MSL level can cause overtemp and unexpected reboot of the router
-if you have one NSR per two LTO drive
>during backups you may got time out if you are using both drives on the same NSR at the same time, and due to HBA time out, you may have it disconnected by the host
Joe Profaizer
Super Advisor

Re: NSR e1200-320 4G and AIX fabric zoning

We have one NSR per two LTO drives. The first NSR actually hosts 2 LTO drives and the robotics.

If timeouts are an issue; how would I go about fixing such?

Thanks for replying; we are dumbfounded so far and can't find a fix for this.
Joe Profaizer
Super Advisor

Re: NSR e1200-320 4G and AIX fabric zoning

I'm still struggling with this issue and have re-posted... Please help if you can.


We have a unique issue with the MSL6060 fibre tape library connectivity with our AIX P5 servers. For whatever reason, the devices go into a DEFINED state once we try to add the tape drives via Data Protector 6.1. We also have had a call logged to IBM regarding this issue but not getting a solution.

All zoning is correct and we have the MSL6060 NSRs utilizing 'Port 0 Device Mapping'. We have no issues with Windows or HP-UX. We also have other AIX server throughout the world working perfectly with other HP Tape Libraries via (non MSL6060s however)fibre connectivity.

Best practice is that we seperate HBAs for data and tape, but we don't have that luxury and we have been through some exhaustive cross-checks, compatibility matrices, etc... and have done hours of trouble shooting to no avail.

Also, all our AIX servers are on the same OS level and TL release. All HBA drivers are the same release, etc...

Has anyone else had a similar issue as described above? If so, please help me out here. ;-)

Thanks in advance.
Andrew Hellyer
Advisor

Re: NSR e1200-320 4G and AIX fabric zoning

I also am having this issue.
I was able to resolve for some time by playing around with the device mapping on the NSR's - but today it stopped, and I am unable to use the tape drives from this host over the fabric.
I can however, see out ESL E-Series (VLS).

I had issues with the VLS also, but I can provide accurate LUN mappings, which seems to work ok.

I am wondering if its the inabilty to manually map a SCSI target/lun to a host on the NSR e1200 that causes this?
Also, I have noticed that the HBA of the AIX server is signed into the brocade switch as an unknown target/initiator, and I am unable to view the driver/firmware levels from the brocade.
What version of AIX are you running, what type of HBA and which firmware?
I believe we are running AIX5.2 with 1.90x13 HBA driver on a emulex LP10000 HBA card.