Created attachment 826223 [details]
ls of /dev/disk/by-path
Description of problem:
If I create two or more virtual host bus adapters on a single physical HBA, only one link gets created in /dev/disk/by-path
I believe this behavior is the result of by-path taking into account different PCI adapters and logical unit serial numbers, but not the scsi host number when creating the links. Since the PCI BDF and the device serial numbers match, only one link gets created.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. create multiple vHBAs on the same physical HBA and expose a target and LUN to the vHBAs
/dev/disk/by-path has only a single link to the device (see attached files)
/dev/disk/by-path has only a link per path to the device
This behavior requires an NPIV capable SAN to reproduce, so please let me know what information I can provide and I'll help in any way I can.
Created attachment 826224 [details]
output of cat /proc/scsi/scsi
Created attachment 826225 [details]
ls of /dev/disk/by-id
Created attachment 826226 [details]
output of virsh nodedev-list --tree (shows the device hierarchy)
I didn't quite finish the description before submitting the BZ:
shows the relationship between the block device and the SCSI device
This request was not resolved in time for the current release.
Red Hat invites you to ask your support representative to
propose this request, if still desired, for consideration in
the next release of Red Hat Enterprise Linux.
I'm working on it right now,
So the by-path format /dev/by-path/pci-xxx-fc-<port_name>-<lun> is ambiguous when there are multiple virtual HBAs on the same physical HBA (NPIV),
I'm not sure about how to properly fix this bug yet, we could add the scsi host number to by-path as Dave suggests or maybe the port_name of the virtual HBA?
I'll try to come up with a solution.
Maurizio, any progress with that patch for path_id?
Michal, see bz 1266934 comment 22
Maurizio, can you please post the patch upstream via GitHub pull request?
Michal, I will post it soon. Sorry for the delay but one customer reported some problems with the first version of the patch and I had to modify it.
They are going to test the new version right now, if they confirm the patch works then I proceed to send it upstream.
I've reverted current patch and merged patch proposed in https://github.com/systemd/systemd/pull/2731
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.