Hello.

I'm trying to update another S8 Sparc system to S10U4. It's a system
with only 2 disks (2×36GB).

askwar@winds03 ~ $ df -k
Filesystem            kbytes    used   avail capacity  Mounted on
/dev/md/dsk/d30      2053605  702341 1289656    36%    /
/dev/md/dsk/d32      3096423 1511244 1523251    50%    /usr
/proc                      0       0       0     0%    /proc
fd                         0       0       0     0%    /dev/fd
mnttab                     0       0       0     0%    /etc/mnttab
/dev/md/dsk/d33      2053605 1305141  686856    66%    /var
swap                 3765864      24 3765840     1%    /var/run
/dev/md/dsk/d34      15483618  610539 14718243     4%    /opt
/dev/md/dsk/d35      9292187 6695410 2596777    73%    /export/home

I'd like to merge /, /usr and /opt. To do so, I removed all the slices or
submirrors or what you want to call it from the existing mirros (like d30
etc.pp.). This means, that one of the disks is now completely unused.
I changed the slice setup to be like this with format:

Part      Tag    Flag     Cylinders         Size            Blocks
  0       root    wm       0 -  3448        4.75GB    (3449/0/0)   9964161
  1        var    wm    3449 -  5082        2.25GB    (1634/0/0)   4720626
  2     backup    wm       0 - 24619       33.92GB    (24620/0/0) 71127180
  3       swap    wm    5083 -  6534        2.00GB    (1452/0/0)   4194828
  4 unassigned    wm       0                0         (0/0/0)            0
  5 unassigned    wm       0                0         (0/0/0)            0
  6        usr    wm    6535 - 24607       24.90GB    (18073/0/0) 52212897
  7 alternates    wm   24608 - 24619       16.93MB    (12/0/0)       34668

I created a "mirror" d0, which only has one side called "d20". s0
is d20. When I'm done, I'll add s0 from the other disk to this
mirror. s6 will be used for ZFS and s7 will be for metadb. I also
created d3 with s1 as d23.

askwar@winds03 ~ $ sudo metainit d20 1 1 c1t1d0s0
d20: Concat/Stripe is setup
askwar@winds03 ~ $ sudo metainit d23 1 1 c1t1d0s3
d23: Concat/Stripe is setup
askwar@winds03 ~ $ sudo metainit d0 -m d20
d0: Mirror is setup
askwar@winds03 ~ $ sudo metainit d3 -m d23
d3: Mirror is setup

Now I'm trying to run lucreate and fail:

askwar@winds03 ~ $ sudo lucreate -n S10 -m /:/dev/md/dsk/d0:ufs -m /var:/dev/md/dsk/d3:ufs -m -:/dev/dsk/c1t1d0s3:swap
Discovering physical storage devices
Discovering logical storage devices
Cross referencing storage devices with boot environment configurations
Determining types of file systems supported
Validating file system requests
Preparing logical storage devices
Preparing physical storage devices
Configuring physical storage devices
Configuring logical storage devices
Analyzing system configuration.
Comparing source boot environment <d30> file systems with the file 
system(s) you specified for the new boot environment. Determining which 
file systems should be in the new boot environment.
ERROR: The following required file systems are not mounted by the currently running OS </export/home>
ERROR: All required PBE file systems are not mounted.

Hm? /export/homes *is* mounted!

askwar@winds03 ~ $ mount | grep /export/home
/export/home on /dev/md/dsk/d35 read/write/setuid/intr/largefiles/logging/noatime/onerror=panic/dev=1540023 on Fri Oct  5 11:52:52 2007
askwar@winds03 ~ $ grep home /etc/vfstab
/dev/md/dsk/d35 /dev/md/rdsk/d35        /export/home ufs     2       yes     largefiles,logging,noatime

Why does lucreate fail?

Alexander

Alexander Skwar wrote:
> Hello.
> 
> I'm trying to update another S8 Sparc system to S10U4. It's a system
> with only 2 disks (2×36GB).
> 
> askwar@winds03 ~ $ df -k
> Filesystem            kbytes    used   avail capacity  Mounted on
> /dev/md/dsk/d30      2053605  702341 1289656    36%    /
> /dev/md/dsk/d32      3096423 1511244 1523251    50%    /usr
> /proc                      0       0       0     0%    /proc
> fd                         0       0       0     0%    /dev/fd
> mnttab                     0       0       0     0%    /etc/mnttab
> /dev/md/dsk/d33      2053605 1305141  686856    66%    /var
> swap                 3765864      24 3765840     1%    /var/run
> /dev/md/dsk/d34      15483618  610539 14718243     4%    /opt
> /dev/md/dsk/d35      9292187 6695410 2596777    73%    /export/home
> 
> I'd like to merge /, /usr and /opt. To do so, I removed all the slices or
> submirrors or what you want to call it from the existing mirros (like d30
> etc.pp.). This means, that one of the disks is now completely unused.
> I changed the slice setup to be like this with format:
> 
> Part      Tag    Flag     Cylinders         Size            Blocks
>   0       root    wm       0 -  3448        4.75GB    (3449/0/0)   9964161
>   1        var    wm    3449 -  5082        2.25GB    (1634/0/0)   4720626
>   2     backup    wm       0 - 24619       33.92GB    (24620/0/0) 71127180
>   3       swap    wm    5083 -  6534        2.00GB    (1452/0/0)   4194828
>   4 unassigned    wm       0                0         (0/0/0)            0
>   5 unassigned    wm       0                0         (0/0/0)            0
>   6        usr    wm    6535 - 24607       24.90GB    (18073/0/0) 52212897
>   7 alternates    wm   24608 - 24619       16.93MB    (12/0/0)       34668
> 
> I created a "mirror" d0, which only has one side called "d20". s0
> is d20. When I'm done, I'll add s0 from the other disk to this
> mirror. s6 will be used for ZFS and s7 will be for metadb. I also
> created d3 with s1 as d23.
> 
> askwar@winds03 ~ $ sudo metainit d20 1 1 c1t1d0s0
> d20: Concat/Stripe is setup
> askwar@winds03 ~ $ sudo metainit d23 1 1 c1t1d0s3
> d23: Concat/Stripe is setup
> askwar@winds03 ~ $ sudo metainit d0 -m d20
> d0: Mirror is setup
> askwar@winds03 ~ $ sudo metainit d3 -m d23
> d3: Mirror is setup
> 
> Now I'm trying to run lucreate and fail:
> 
> askwar@winds03 ~ $ sudo lucreate -n S10 -m /:/dev/md/dsk/d0:ufs -m /var:/dev/md/dsk/d3:ufs -m -:/dev/dsk/c1t1d0s3:swap
> Discovering physical storage devices
> Discovering logical storage devices
> Cross referencing storage devices with boot environment configurations
> Determining types of file systems supported
> Validating file system requests
> Preparing logical storage devices
> Preparing physical storage devices
> Configuring physical storage devices
> Configuring logical storage devices
> Analyzing system configuration.
> Comparing source boot environment <d30> file systems with the file 
> system(s) you specified for the new boot environment. Determining which 
> file systems should be in the new boot environment.
> ERROR: The following required file systems are not mounted by the currently running OS </export/home>
> ERROR: All required PBE file systems are not mounted.
> 
> Hm? /export/homes *is* mounted!
> 
> askwar@winds03 ~ $ mount | grep /export/home
> /export/home on /dev/md/dsk/d35 read/write/setuid/intr/largefiles/logging/noatime/onerror=panic/dev=1540023 on Fri Oct  5 11:52:52 2007
> askwar@winds03 ~ $ grep home /etc/vfstab
> /dev/md/dsk/d35 /dev/md/rdsk/d35        /export/home ufs     2       yes     largefiles,logging,noatime
> 
> Why does lucreate fail?
> 
> Alexander

I've seen this weirdness during one of the test U3->U4 upgrades.  It's a bug in lucreate, maybe rarely triggered by a particular environment.  The root cause is an illegal ERE pattern that lucreate constructs for egrep(1) from a mount(1m) output to use against the list of mount points from /etc/vfstab.  For the very last FS reported by mount(1m), the pattern for its mount point is broken, so it never matches the right entry in /etc/vfstab creating an illusion that one of the FS listed in /etc/vfstab is currently not mounted.  In my case it was /opt/local, in yours it's /export home.

I found at least two workarounds, there may be more, of course.

First, (indirectly) modify mnttab so that none of the actually used by the current BE filesystems goes last in the mount(1m) output, e.g. by [making and] mounting some extra unneeded/fake FS.

Second (good only if the offending filesystem is shared), comment out the unlucky FS entry in /etc/vfstab, run lucreate, reinstate /etc/vfstab and add by hand the missing FS entry to the appropriate /etc/lu/ICF.? (/usr/lib/lu/lumk_iconf can be used to generate the exact line) before running luupgrade.


-- 
/ynp