multipath-tools-0.4.8-r2/r1 needs SYSFS_DEPRECATED_V2 in kernel config which cause error at boot time : "udev: deprecated sysfs layout; update the kernel or disable CONFIG_SYSFS_DEPRECATED; some udev features will not work correctly" (2.6.28-r9) Without deprecated sysfs : alpha linux # multipath -ll domsU2 (3600a0b80001f3e49000055764ad8217b) dm-0 , [size=160G][features=0][hwhandler=0] \_ round-robin 0 [prio=1][enabled] \_ #:#:#:# sda 8:0 [active][ready] \_ round-robin 0 [prio=3][active] \_ #:#:#:# sdd 8:48 [active][ready] With : domsU2 (3600a0b80001f3e49000055764ad8217b) dm-0 IBM ,1722-600 [size=160G][features=1 queue_if_no_path][hwhandler=1 rdac] \_ round-robin 0 [prio=0][enabled] \_ 2:0:0:0 sda 8:0 [active][ghost] \_ round-robin 0 [prio=3][active] \_ 3:0:0:0 sdd 8:48 [active][ready] Perharps the sams as this one : https://bugs.launchpad.net/ubuntu/+source/multipath-tools/+bug/307032
that udev output is not an error, it's a warning. it doesnt break because of it.
I know... It's just that today multipath-tools should not use the sysfs deprecated part ;)
Is there a 1-on-1 equivalent as for where the new SysFS entries may be found so as a patch to be provided? I can take a look at the sources to do that but I'll probably need a little help.
I encounter a similar issue recently when I disabled the deprecated option in the kernel. I did some looking around and tried the latest version in portage (multipath-tools-0.4.8-r20 and got the same results with it not properly seeing our MD3000i. Next, I made a temp ebuild and pointed it to the git repo and built the latest from git. And lo and behold everything is working properly again. I noticed that Fedora is using a snapshot of git with other patches so it might be a good idea if Gentoo went a similar route to at least get our package more up to date. It doesn't seem that upstream is tagging any new releases recently. What are your thoughts on creating a newer ebuild based on a snapshot from git? Thanks-
i've been wondering about a newer one based on git for a while, but was trying to encourage upstream to just make a release instead.
(In reply to comment #5) > i've been wondering about a newer one based on git for a while, but was trying > to encourage upstream to just make a release instead. > I'd say if upstream hasn't responded in a reasonable time, go for it but I agree upstream making a release is a better option. This is a blocker for me in this particular case so either I need a proper fix or just a temp one in our overlay.
what git revision snapshot did you use for your deployment? I'm trying to see about picking a good one.
(In reply to comment #7) > what git revision snapshot did you use for your deployment? I'm trying to see > about picking a good one. > The version I pulled was b68400c but it was just the latest I pulled at the time. Feel free to use that but I didn't do a lot of testing. Thanks-
Bug 377709 removed sysfsutils depend from 0.4.8 and 0.4.9 wrt this upstream commit: http://git.kernel.org/gitweb.cgi?p=linux/storage/multipath-tools/.git;a=commit;h=8ccef7c766a4891140488d12d93e5eb930271bf2 I would have expected any SYSFS_DEPRECATED_V2 requiring code to die with it. Is this still a problem with 0.4.9 series?
Should be fine with 0.4.9 series.