[arch-general] dmraid-1.0.0rc15 in testing, need you help on this!

[David C. Rankin]

On Sunday 21 June 2009 02:07:53 am Tobias Powalowski wrote:
> Am Sonntag 21 Juni 2009 schrieb David C. Rankin:
> > On Friday 19 June 2009 04:17:45 pm Tobias Powalowski wrote:
> > > Hi guys,
> > > I put again a 1.0.0rc15 package to testing, i can't test it myself so i
> > > need you to help me on this.
> > >
> > > Is it true that name scheme changed from:
> > > name1 to namep1 for partitions?
> > > And how are your arrays named?
> > > Are your arrays assembled correct?
> > > Does everything work as before?
> > > Thanks for helping me on this, it will improve our dmraid support.
> > >
> > > greetings
> > > tpowa
> >
> > Tobias,
> >
> > 	I have installed 1.0.0rc15 but I haven't rebooted yet. Are you saying I
> > need to put the 'p' here:
> >
> > # (0) Arch Linux
> > title  Arch Linux
> > root (hd1,5)
> > kernel /vmlinuz26 root=/dev/mapper/nvidia_ecaejfdi<p>5 ro vga=0x31a
> > initrd /kernel26.img
> >
> > # (1) Arch Linux
> > title  Arch Linux Fallback
> > root (hd1,5)
> > kernel /vmlinuz26 root=/dev/mapper/nvidia_ecaejfdi<p>5 ro
> > initrd /kernel26-fallback.img
> >
> > # (2) SuSE Linux 11.0
> > title openSuSE Linux 11.0
> > rootnoverify (hd0,4)
> > makeactive
> >
> > # (3) memtest86+
> > title  Memtest86+  [/memtest86+/memtest.bin]
> > kernel (hd1,5)/memtest86+/memtest.bin
> >
> > AND in fstab
> >
> > /dev/mapper/nvidia_ecaejfdi<p>5 / ext3 defaults 0 1
> > /dev/mapper/nvidia_ecaejfdi<p>6 /boot ext3 defaults 0 1
> > /dev/mapper/nvidia_ecaejfdi<p>7 /home ext3 defaults 0 1
> > /dev/mapper/nvidia_ecaejfdi<p>8 swap swap defaults 0 0
>
> Yes this might happen, last time i bumped to this version, people reported
> this.
> I'm not sure if everyone needs it though, i need feedback on this.
> greetgins
> tpowa

Tobias:

	Yes, it DID happen. The post-install warning was correct in my case. I 
modified both /boot/grub/menu.lst and /etc/fstab to include the atavistic 'p' 
before the number within the device mapper label and Arch booted just fine (of 
course I got the black bean on having fsck check part three which was an 
enjoyable break while it completed. Everything seems to be humming away just 
as it was before except for the new 'p'.

01:48 archangel:~> sudo dmraid -r
/dev/sdd: nvidia, "nvidia_ecaejfdi", mirror, ok, 1465149166 sectors, data@ 0
/dev/sdc: nvidia, "nvidia_fdaacfde", mirror, ok, 976773166 sectors, data@ 0
/dev/sdb: nvidia, "nvidia_ecaejfdi", mirror, ok, 1465149166 sectors, data@ 0
/dev/sda: nvidia, "nvidia_fdaacfde", mirror, ok, 976773166 sectors, data@ 0

[01:52 archangel:/home/david] # dmraid -s -s
*** Active Set
name   : nvidia_ecaejfdi
size   : 1465149056
stride : 128
type   : mirror
status : ok
subsets: 0
devs   : 2
spares : 0
*** Active Set
name   : nvidia_fdaacfde
size   : 976773120
stride : 128
type   : mirror
status : ok
subsets: 0
devs   : 2
spares : 0


	One Question: "Why is the 'p' necessary?" It breaks all the scripts I had 
that deal with mounting and unmounting the arrays for the SuSE install when I 
am running Arch. SuSE is on a separate array on the same machine. (SuSE is on 
a dm array created out of sda & sdc and Arch is on an array created on sdb & 
sdd) 

	Now when I boot into Arch and need to mount/unmount the arrays from the SuSE 
install, they fail, because the scripts have no p... Arch now boots like this:

01:18 archangel:~> ls -1 /dev/mapper
control
nvidia_ecaejfdi
nvidia_ecaejfdip5
nvidia_ecaejfdip6
nvidia_ecaejfdip7
nvidia_ecaejfdip8
nvidia_fdaacfde
nvidia_fdaacfdep5
nvidia_fdaacfdep6
nvidia_fdaacfdep7
nvidia_fdaacfdep8

	At least when SuSE boots it still has my old dm labels so the scripts on SuSE 
still work to manipulate mounting and unmount of the arrays the inactive Arch 
is on:

01:21 ecstasy:~> ls -1 /dev/mapper
control
nvidia_ecaejfdi
nvidia_ecaejfdi_part1
nvidia_ecaejfdi_part5
nvidia_ecaejfdi_part6
nvidia_ecaejfdi_part7
nvidia_ecaejfdi_part8
nvidia_fdaacfde
nvidia_fdaacfde_part1
nvidia_fdaacfde_part5
nvidia_fdaacfde_part6
nvidia_fdaacfde_part7
nvidia_fdaacfde_part8

	So I guess I will have to edit all the scripts to put an extra 'p' in them as 
well. (Really kind of a pain, but if this new scheme is going to be the 
standard, I can do it, ...but if this is just a 'proposed' change, then I 
would rather not.

	So is this Official? If so I don't mind changing them around and I know SuSE 
will eventually catch up. I've cc'ed Heinz Mauelshagen (the dev at Redhat to 
see if he thinks this is permanent. This does present a problem for Linux dual 
boot boxes. What was the reason to the change anyway?

	Tobias, let me know if you want me to run any more test. (I don't have 
anything to rebuild right now or I would try the -R command which is an 
EXCELLENT addition to dmraid. Kuddos to the brainchild behind that.

-- 
David C. Rankin, J.D.,P.E.
Rankin Law Firm, PLLC
510 Ochiltree Street
Nacogdoches, Texas 75961
Telephone: (936) 715-9333
Facsimile: (936) 715-9339
www.rankinlawfirm.com