importnace of scsi_transport_fc in scsi_mod module

[skt_skt]

Enterprise Linux Enterprise Linux Server release 5.1 (Carthage)

2.6.18-53.1.21.2.1.el5 #1 SMP Thu Nov 13 23:56:36 EST 2008 x86_64 x86_64 x86_64 GNU/Linux

# lsmod|grep -i qla
qla2400 238592 0
qla2xxx 310432 49 qla2400
qla2xxx_conf 335240 1
intermodule 37764 2 qla2xxx,qla2xxx_conf
scsi_mod 186361 7 emcp,qla2xxx,sg,scsi_transport_fc,libata,aacraid,sd_mod

I am having trouble discovering(seven discovered and there is more than ten left ) the LUNS presented from EMC(DMX-3)array and noticed the working servers have scsi_transport_fc under scsi_mod.

Any one know what scsi_transport_fc does here exactly?

[skt_skt]

The hosts are able to see the LUNs with combinations(VBUS+Taget+Lun from maskign point of view ) up to [0 0F 007] , not starting from [1 00 000]. Could you help us in explaining that?

Any one know what would prevent Linux to see them? Any one had a simialr problem??

[Matti_Kurkela · Accepted solution]

scsi_transport_fc is just a module that understands FibreChannel-specific attributes. In earlier versions, this code was embedded in each SCSI/FC driver, but now it has been made into a separate module to avoid code duplication (no need to fix the same bugs in two or more different places).

If the non-working servers don't have it, that would sort of indicate that their FC drivers (or perhaps the kernel) are older than on the hosts using the scsi_transport_fc module. Some updates might be needed.

For your LUN discovery problem, I'd start by logging onto EMC Powerlink website (powerlink.emc.com, EMC support identifier required to register) and running a search on their compatibility database, listing the array, HBA and OS as search terms. If the search results include some HBA firmware or storage configuration requirements, I would check them very carefully and update as necessary.

MK