Bug 63806

Summary:	udev 202 mangles persistent disk/by-path w/ Intel C606 Dual 4-port SAS
Product:	systemd	Reporter:	me
Component:	general	Assignee:	systemd-bugs
Status:	RESOLVED FIXED	QA Contact:	systemd-bugs
Severity:	normal
Priority:	medium	CC:	freedesktop-bugs, systemd
Version:	unspecified
Hardware:	Other
OS:	All
See Also:	http://bugs.gentoo.org/show_bug.cgi?id=466614
Whiteboard:
i915 platform:		i915 features:

Description me 2013-04-22 15:13:29 UTC

by-path persistent device names overlap the 2nd 4-port section of the controller on top of the first:

pci-0000:03:00.0-sas-0x500304800e246e00-lun-0 -> ../../sdc
pci-0000:03:00.0-sas-0x500304800e246e00-lun-0-part1 -> ../../sdc1
pci-0000:03:00.0-sas-0x500304800e246e01-lun-0 -> ../../sdd
pci-0000:03:00.0-sas-0x500304800e246e01-lun-0-part1 -> ../../sdd1
pci-0000:03:00.0-sas-0x500304800e246e01-lun-0-part2 -> ../../sdg2
pci-0000:03:00.0-sas-0x500304800e246e02-lun-0 -> ../../sde
pci-0000:03:00.0-sas-0x500304800e246e02-lun-0-part1 -> ../../sde1



sd[cde] are on ports 6:0-3 and sd[fg] are on ports 7:0-1 where 6 is the first half and 7 is the second half of the "dual" hardware config.  Where a link already exists with a particular name, no new one is created to the 2nd half.  This is why /dev/sdf is not linked to and why there is an sdg2 but no sdg or sdg1 (they would be the same as sdd and sdd1).

/devices/pci0000:00/0000:00:01.0/0000:01:00.0/0000:02:08.0/0000:03:00.0/host6/port-6:1/end_device-6:1/target6:0:1/6:0:1:0/block/sdd

/devices/pci0000:00/0000:00:01.0/0000:01:00.0/0000:02:08.0/0000:03:00.0/host7/port-7:1/end_device-7:1/target7:0:1/7:0:1:0/block/sdg

E: DEVLINKS=/dev/disk/by-id/ata-M4-CT128M4SSD2_00000000113003152D83 /dev/disk/by-id/wwn-0x500a075103152d83 /dev/disk/by-path/pci-0000:03:00.0-sas-0x500304800e246e01-lun-0

E: DEVLINKS=/dev/disk/by-id/ata-WD_WD30EFRX-68AX9N0_WD-WMC1T1636965 /dev/disk/by-id/wwn-0x50014ee0ae28a327 /dev/disk/by-path/pci-0000:03:00.0-sas-0x500304800e246e01-lun-0



Whatever is determining ID_PATH isn't doing it right in this case.  Let me know what additional information I can provide.

Comment 1 Kay Sievers 2013-09-12 18:28:06 UTC

This is not really maintained, because none of us understands or uses SAS
hardware.

The people involved in the original code somehow can't agree on how to do this,
and changed the way these links are composed, but it seems it was never solved
sufficiently.

Sorry, I'm just closing this, as none of us will or can work on fixing it.

The longer-term plan for udev is to separate-out all the non-trivial SCSI
things and leave it to a separate package maintained by somebody else who
is willing and capable to do that. None of us is in the position to do that.

Comment 2 Zbigniew Jedrzejewski-Szmek 2014-12-05 13:16:56 UTC

The way that SAS by-path ids are calculated was changed with commit
http://cgit.freedesktop.org/systemd/systemd/commit/?id=66bba0e701. It should also fix this problem, please check!

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.