current state of (F11) installation

Sun Aug 16 18:59:53 UTC 2009

Hi all,

under the following prerequisites one specific installation path of F11
(taken from Karsten's pungi tree dated 20080804) now works until the
shutdown of the installation system, i.e. when anaconda passes control
back to linuxrc.s390:
- all DASDs offline (I have a non-existant devno in conf file)
- activate one zFCP-attached SCSI disk from within parm/conf file
- partition a small swap partition and the rest of /dev/sda as rootfs
  with ext3
- on top of Karsten's already modified
  anaconda-11.5.0.59-1.fc11.kh.1.s390x, you need the following patches
  I posted to anaconda-devel (and most of which are also upstream):
  * for initrd.img:
    * suggest disabled X11-forwarding for ssh login in linuxrc.s390
    * make IPv4 configuration in linuxrc.s390 compatible with NM in loader
    * Handle activation of DASDs in linuxrc.s390 since loader no longer
works
    * Correctly inform the user once about obsolete parm/conf file
options on s390
  * for updates.img:
    * correctly activate zFCP LUN on s390
    * correctly deactivate zFCP LUN on s390
    * prevent getting started up or shutdown again while already in such
state
    * Log zFCP errors in addition to displaying dialog box

Adding the one SCSI disk interactively in Anaconda works, but
iw/autopart_type.py:self.bootcombo doesn't get updated and thus reports
an error "Must select a drive to use as ""the bootable device." and
won't continue. This whole mechanism including the UI dropdown box is
meaningless and even misleading for s390x since the zipl bootloader
always has to go on the disk device where /boot lives.

At the end of anaconda, when packages and the bootloader have been
installed, the reboot doesn't work because /sbin/shutdown is no longer
in initrd.img:

> [2009-08-16 20:22:05,812]  WARNING: Storage.writeKS not completely implemented 
> 00: HCPGSP2629I The virtual machine is placed in CP mode due to a SIGP stop from
>  CPU 02.
> 01: HCPGSP2629I The virtual machine is placed in CP mode due to a SIGP stop from
>  CPU 02.
> /init: line 92: /sbin/shutdown: No such file or directory 
> /init: line 92: exec: /sbin/shutdown: cannot execute: No such file or directory 
> Kernel panic - not syncing: Attempted to kill init! 
>  Kernel BUG at 0000000000000002 [verbose debug info unavailable] 
> illegal operation: 0001 [#1] SMP  
> Modules linked in: xfs exportfs vfat fat ext2 zfcp scsi_transport_fc scsi_tgt xt
> s lrw gf128mul sha256_generic cbc dm_crypt dm_round_robin dm_multipath linear ra
> id10 raid456 async_xor async_memcpy async_tx xor raid1 raid0 dasd_fba_mod dasd_d
> iag_mod iscsi_tcp libiscsi_tcp libiscsi scsi_transport_iscsi squashfs cramfs qet
> h_l3 ipv6 qeth qdio ccwgroup dasd_eckd_mod dasd_mod 
> CPU: 2 Not tainted 2.6.29.5-191.fc11.s390x #1 
> Process init (pid: 1, task: 000000003fe98000, ksp: 000000003fe97c58) 
> Krnl PSW : (null) 0000000000000002 (0x2) 
>            R:0 T:0 IO:0 EX:0 Key:0 M:0 W:0 P:0 AS:0 CC:0 PM:0 EA:0 
> Krnl GPRS: 0000000000000000 0000000000001000 000000003fe97c90 0000000000000000 
>            00000000006aa62e ffffffff00000001 0000000000000000 00000000000003e0 
>            000000000005793c 00000000006aa62e 0000000000000000 000000003fe97be0 
>            0000000000634b88 0000000000419608 0000000000023e5e 000000003fe97be0 
> Krnl Code:>0000000000000002: 0000               unknown 
>            0000000000000004: 0000               unknown 
>            0000000000000006: 0000               unknown 
>            0000000000000008: 0000               unknown 
>            000000000000000a: 0000               unknown 
>            000000000000000c: 0000               unknown 
>            000000000000000e: 0000               unknown 
>            0000000000000010: 0000               unknown 
> Call Trace: 
> ([<0000000000000040>] 0x40) 
>  [<0000000000023e5e>] on_panic_notify+0x52/0x70 
>  [<0000000000415708>] notifier_call_chain+0x48/0x9c 
>  [<00000000004157ca>] atomic_notifier_call_chain+0x32/0x44 
>  [<000000000040f4c6>] panic+0xd6/0x204 
>  [<0000000000057070>] do_exit+0xa8/0x884 
>  [<000000000005790e>] do_group_exit+0xc2/0xf0 
>  [<000000000005798a>] SyS_exit_group+0x4e/0x60 
>  [<0000000000029086>] sysc_noemu+0x10/0x16 
>  [<000002000010e20a>] 0x2000010e20a 
> Last Breaking-Event-Address: 
>  [<0000000000023dd4>] stop_run+0xec/0x124 
>  <4>---[ end trace fc7da2acd33fb952 ]--- 
> Fixing recursive fault but reboot is needed! 

You can manually trigger the reboot using the magic sysrq key sequence:
^-b
The VM guest then reboots (reIPL) correctly:

> 00: HCPGSP2629I The virtual machine is placed in CP mode due to a SIGP stop from
>  CPU 02.
> 01: HCPGSP2629I The virtual machine is placed in CP mode due to a SIGP stop from
>  CPU 02.
> 02: HCPLDI2816I Acquiring the machine loader from the processor controller.
> 02: HCPLDI2817I Load completed from the processor controller. 
> 02: HCPLDI2817I Now starting the machine loader.
> 00: HCPGSP2630I The virtual machine is placed in CP mode due to a SIGP stop and 
> store status from CPU 02.
> 01: HCPGSP2630I The virtual machine is placed in CP mode due to a SIGP stop and 
> store status from CPU 02.
> 02: MLOEVL012I: Machine loader up and running (version v1.4).
> 02: MLOPDM003I: Machine loader finished, moving data to final storage location.
> Initializing cgroup subsys cpuset 
> Initializing cgroup subsys cpu 
> Linux version 2.6.29.5-191.fc11.s390x (mockbuild at zedora3.z900.redhat.com) (gcc v
> ersion 4.4.0 20090506 (Red Hat 4.4.0-4) (GCC) ) #1 SMP Wed Jun 24 15:02:47 EDT 2
> 009 
> setup: Linux is running as a z/VM guest operating system in 64-bit mode 
> setup: Address spaces switched, mvcos available 
> Zone PFN ranges: 
>   DMA      0x00000000 -> 0x00080000 
>   Normal   0x00080000 -> 0x00080000 
> Movable zone start PFN for each node 
> early_node_map[1] active PFN ranges 
>     0: 0x00000000 -> 0x00040000 
> Built 1 zonelists in Zone order, mobility grouping on.  Total pages: 258560 
> Kernel command line: root=UUID=8ee05092-a8e0-4081-8076-0a9dec7ea2fc  
> ...
> Freeing unused kernel memory: 224k freed 
> Mounting proc filesystem 
> Mounting sysfs filesystem 
> Creating /dev 
> Creating initial device nodes 
> scsi0 : zfcp 
> scsi1 : zfcp 
> qdio: 0.0.3c1b ZFCP on SC 14 using AI:1 QEBSM:1 PCI:1 TDD:1 SIGA: W AO  

However, the initrd code seems to come to a halt when activating the
SCSI disk with the rootfs. Since mkinitrd is replaced with dracut for
F12, IMHO it doesn't make much sense trying to fix mkinitrd.

On using one DASD instead of a SCSI disk, I get errors when entering the
partitioner about unfulfilled constraints of parted. I suspect this to
be due to DASD support in parted, since SCSI worked as stated above.

Steffen

Linux on System z Development

IBM Deutschland Research & Development GmbH
Vorsitzender des Aufsichtsrats: Martin Jetter
Geschäftsführung: Erich Baier
Sitz der Gesellschaft: Böblingen
Registergericht: Amtsgericht Stuttgart, HRB 243294