Bug#575682: marked as done (installation-reports: successful installation on hercules-emulated s390x)
Your message dated Tue, 30 Mar 2010 06:00:09 +0200
with message-id <firstname.lastname@example.org>
and subject line Re: Bug#575682: installation-reports: successful installation on hercules-emulated s390x
has caused the Debian Bug report #575682,
regarding installation-reports: successful installation on hercules-emulated s390x
to be marked as done.
This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.
(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact email@example.com
Debian Bug Tracking System
Contact firstname.lastname@example.org with problems
--- Begin Message ---
Boot method: card reader
Image version: http://people.debian.org/~fjp/d-i/s390/images/daily/generic/
Date: Thu, 25 Mar 2010 15:11:54 -0500
Machine: Fake z/Arch, emulated using hercules 3.07-1
setup described at http://www.josefsipek.net/docs/s390-linux/hercules-s390.html
Actual machine: Dell Inspiron 4000 (a laptop with a Pentium III in it)
$ df -Tl | grep '/dev'
/dev/dasda1 ext4 2366164 621636 1624332 28% /
/dev/dasdb1 ext4 2366164 69088 2176880 4% /home
Using whole disks. For example:
Disk /dev/dasda doesn't contain a valid partition table
Note: sector size is 4096 (not 512)
Disk /dev/dasda: 2461 MB, 2461777920 bytes
15 heads, 12 sectors/track, 3339 cylinders
Units = cylinders of 180 * 4096 = 737280 bytes
Disk identifier: 0xe5e5e5e5
Base System Installation Checklist:
[O] = OK, [E] = Error (please elaborate below), [ ] = didn't try it
Initial boot: [O]
Detect network card: [O]
Configure network: [O]
Detect CD: [ ]
Load installer modules: [O]
Detect hard drives: [O]
Partition hard drives: [O]
Install base system: [O]
Clock/timezone setup: [O]
User/password setup: [O]
Install tasks: [O]
Install boot loader: [O]
Overall install: [O]
1. This is my first successful direct squeeze install on s390. I tried a
daily build in February (the daily builds weren't being updated at the
time, so I assume this was actually a January, 2010 daily build) and
HHCCP007I CPU0000 architecture mode set to ESA/390
HHCCP011I CPU0000: Disabled wait state
and did nothing more. So I was excited to see that this time I received
HHCCP007I CPU0000 architecture mode set to z/Arch
which seems to mean the SIGP worked.
The rest of this installation report will be boring: because I have my
notes, I am pointing out the steps that were noticeably slow without a
progress display. Everything worked great. :)
2. Generating an SSH host key takes a long time. I wish there was a way
to bypass this step and use telnet instead.
I ssh'd in with ssh -c blowfish and chose "Start installer" for a simple
install. The location questions were as I'd expect.
3. Somewhat slow to show the "Choose mirror" list.
4. Somewhat slow after downloading release.
5. As Josef Sipek notes, the installer calls DASDs "Disk access storage
devices". That should be "Direct access".
6. "Disks need to be formatted before you can create partitions": it
isn't clear to me what this means --- partman coped fine without
this step. Maybe it would be clearer to be more explicit:
"Reinitialize partition table?" or something similar would have
been much clearer (if that is what was meant), though perhaps too
7. Blue screen for a while after "Starting up the partitioner" finishes
and before I can interact with the partition manager.
8. The partitioner is even nicer than I remember it. I went with ext4
for both partitions (might have used btrfs if that was available).
9. It asks if I want to go back and use swap. IIRC to avoid fragmentation
over time it is really necessary to have swap. Is there a suggested
way to deal with this in emulated environments like this one, where
swapping to memory would be by far preferrable over doing disk I/O?
Anyway, I said "no swap" in keeping with Josef's instructions.
10. Installing the base system takes a while.
I hope there is just a final interview after that before I can walk
away and let the installer do its thing, but that is not to be.
man-db triggers take a lot of time. Maybe they could be saved up
for the end of the install?
The progress messages often have the form "Retrieving 1 file of 2".
This is not very useful, since it doesn't say 1 file of 2 in what
Incidentally, an installer timechart would be kind of neat to play
11. Now comes an interview to set up root and the first user. There
was no option to set up sudo here. (I like to use a dedicated
"sudoer" user to take advantage of sudo caching credentials but
timing out appropriately without making my main account so easily
root-equivalent while I am carrying out administrative tasks.)
12. Configuring APT takes a while.
13. Then "Select and install software" --- this is the bulk of the
task. I think man-db triggers make this slower than it needs
14. At progress=13%, the popcon dialog comes up. Couldn't it have
15. Eventually it starts up tasksel. Not much to say about that ---
it works pretty well. I chose to install the standard system
16. For the initial boot:
- starting up the login shell takes a _long_ time (completion
- -bash: no job control in this shell
- C locale (no non-ASCII characters). Not sure why.
17. Something strange happened when I suspended reportbug: it seems
to have accepted a ^C that was meant to be directed at bash.
Please select a severity level: [normal]
+ Stopped reportbug installation-reports
jrn@s390:~$ ls /usr/share/doc/insta^C
jrn@s390:~$ ls /usr/share/bug/installation-reportat^C
+ Terminated reportbug installation-reports
jrn@s390:~$ ls /usr/share/bug/installation-reports
jrn@s390:~$ ls /usr/share/bug/installation-reports/script
^[[Ajrn@s390:~$ less /usr/share/bug/installation-reports/script
jrn@s390:~$ less /usr/share/bug/installation-report/script
-bash: fg: current: no such job
18. Next step is to install telnet as an inetd service to get rid of some
So, very smooth. I’m impressed.
DISTRIB_DESCRIPTION="Debian GNU/Linux installer"
DISTRIB_RELEASE="6.0 (squeeze) - installer build 20100325-05:20"
uname -a: Linux s390 2.6.32-3-s390x #1 SMP Thu Feb 25 15:07:55 UTC 2010 s390x GNU/Linux
lspci -knn: pcilib: Cannot open /proc/bus/pci
lspci -knn: lspci: Cannot find any working access method.
lsmod: Module Size Used by
lsmod: dm_mod 82061 0
lsmod: md_mod 100992 0
lsmod: xfs 615180 0
lsmod: exportfs 4658 1 xfs
lsmod: ext4 420316 2
lsmod: jbd2 108020 1 ext4
lsmod: crc16 1471 1 ext4
lsmod: ext3 154425 0
lsmod: jbd 70570 1 ext3
lsmod: vfat 11890 0
lsmod: fat 61179 1 vfat
lsmod: nls_base 8049 2 vfat,fat
lsmod: ext2 80475 0
lsmod: mbcache 8226 3 ext4,ext3,ext2
lsmod: ctcm 79453 0
lsmod: fsm 3674 1 ctcm
lsmod: cu3088 2368 1 ctcm
lsmod: dasd_eckd_mod 69042 8
lsmod: ccwgroup 7786 1 cu3088
lsmod: dasd_mod 63773 5 dasd_eckd_mod
df: Filesystem 1K-blocks Used Available Use% Mounted on
df: tmpfs 124944 20 124924 0% /dev
df: /dev/dasda1 2366164 621224 1624744 28% /target
df: /dev/dasdb1 2366164 69080 2176888 3% /target/home
df: /dev/dasda1 2366164 621224 1624744 28% /dev/.static/dev
df: tmpfs 124944 20 124924 0% /target/dev
free: total used free shared buffers
free: Mem: 249888 222908 26980 0 24420
free: Swap: 0 0 0
free: Total: 249888 222908 26980
/proc/cmdline: ro locale=C
/proc/cpuinfo: vendor_id : IBM/S390
/proc/cpuinfo: # processors : 1
/proc/cpuinfo: bogomips per cpu: 171.00
/proc/cpuinfo: features : esan3 zarch stfle msa ldisp eimm etf3eh highgprs
/proc/cpuinfo: processor 0: version = 00, identification = 000069, machine = 9672
/proc/iomem: 00000000-0fffffff : System RAM
/proc/iomem: 00000000-0035ba37 : Kernel code
/proc/iomem: 0035ba38-0052365f : Kernel data
/proc/interrupts: EXT: 955097
/proc/interrupts: I/O: 168569
/proc/meminfo: MemTotal: 249888 kB
/proc/meminfo: MemFree: 26980 kB
/proc/meminfo: Buffers: 24428 kB
/proc/meminfo: Cached: 167084 kB
/proc/meminfo: SwapCached: 0 kB
/proc/meminfo: Active: 69028 kB
/proc/meminfo: Inactive: 107024 kB
/proc/meminfo: Active(anon): 2016 kB
/proc/meminfo: Inactive(anon): 3920 kB
/proc/meminfo: Active(file): 67012 kB
/proc/meminfo: Inactive(file): 103104 kB
/proc/meminfo: Unevictable: 21376 kB
/proc/meminfo: Mlocked: 0 kB
/proc/meminfo: SwapTotal: 0 kB
/proc/meminfo: SwapFree: 0 kB
/proc/meminfo: Dirty: 368 kB
/proc/meminfo: Writeback: 0 kB
/proc/meminfo: AnonPages: 5952 kB
/proc/meminfo: Mapped: 3184 kB
/proc/meminfo: Shmem: 20 kB
/proc/meminfo: Slab: 19064 kB
/proc/meminfo: SReclaimable: 15800 kB
/proc/meminfo: SUnreclaim: 3264 kB
/proc/meminfo: KernelStack: 1424 kB
/proc/meminfo: PageTables: 256 kB
/proc/meminfo: NFS_Unstable: 0 kB
/proc/meminfo: Bounce: 0 kB
/proc/meminfo: WritebackTmp: 0 kB
/proc/meminfo: CommitLimit: 124944 kB
/proc/meminfo: Committed_AS: 10672 kB
/proc/meminfo: VmallocTotal: 1048576 kB
/proc/meminfo: VmallocUsed: 4088 kB
/proc/meminfo: VmallocChunk: 1039572 kB
/proc/meminfo: HugePages_Total: 0
/proc/meminfo: HugePages_Free: 0
/proc/meminfo: HugePages_Rsvd: 0
/proc/meminfo: HugePages_Surp: 0
/proc/meminfo: Hugepagesize: 1024 kB
-- System Information:
Debian Release: squeeze/sid
APT prefers testing
APT policy: (500, 'testing')
Architecture: s390 (s390x)
Kernel: Linux 2.6.32-3-s390x (SMP w/1 CPU core)
Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968)
Shell: /bin/sh linked to /bin/dash
--- End Message ---
--- Begin Message ---
Thanks for the report Jonathan.
On Sunday 28 March 2010, Jonathan Nieder wrote:
> The rest of this installation report will be boring: because I have my
> notes, I am pointing out the steps that were noticeably slow without a
> progress display. Everything worked great. :)
You have to realize that this is about the slowest platform you can get for
Debian Installer. So yes, there are points where things can get annoying,
but as long as that problem does not exist for installs on more "regular"
systems, I doubt we'll spend much effort on "fixing" it.
> 2. Generating an SSH host key takes a long time. I wish there was a way
> to bypass this step and use telnet instead.
telnet is inherently insecure; as the issue is largely caused by the
slowness of the platform, this won't be changed.
> 5. As Josef Sipek notes, the installer calls DASDs "Disk access storage
> devices". That should be "Direct access".
Thanks. I've fixed that.
> 6. "Disks need to be formatted before you can create partitions": it
> isn't clear to me what this means --- partman coped fine without
> this step. Maybe it would be clearer to be more explicit:
> "Reinitialize partition table?" or something similar would have
> been much clearer (if that is what was meant), though perhaps too
> jargon heavy.
No. It really is low-level formatting (it takes ages). This is more your
unfamiliarity with the platform than anything that should be improved IMO.
> 9. It asks if I want to go back and use swap. IIRC to avoid
> fragmentation over time it is really necessary to have swap. Is there a
> suggested way to deal with this in emulated environments like this one,
> where swapping to memory would be by far preferrable over doing disk
> I/O? Anyway, I said "no swap" in keeping with Josef's instructions.
I've always set up swap on my Hercules systems, but I don't really have an
answer for you.
> 10. Installing the base system takes a while.
> man-db triggers take a lot of time. Maybe they could be saved up
> for the end of the install?
man-db updates are already implemented as triggers. We really don't want to
mess about too much with standard package management in the installer.
> The progress messages often have the form "Retrieving 1 file of 2".
> This is not very useful, since it doesn't say 1 file of 2 in what
The info comes from APT. I agree it's not perfect, but on slower network
connections it does provide some additional progress info. Possibly we
will improve this at some point, but that may require an extention of the
> 11. Now comes an interview to set up root and the first user. There
> was no option to set up sudo here. (I like to use a dedicated
> "sudoer" user to take advantage of sudo caching credentials but
> timing out appropriately without making my main account so easily
> root-equivalent while I am carrying out administrative tasks.)
There is an option to setup sudo, but only when installing at medium or low
debconf priority. You can set medium debconf priority using the
appropriate option in the main menu, or by adding 'priority=medium' in
> 14. At progress=13%, the popcon dialog comes up. Couldn't it have
> asked sooner?
Short answer: no. The dialog is not displayed by the Installer, but by the
> 16. For the initial boot:
> - starting up the login shell takes a _long_ time (completion
> script cruft)?
Yes. Uninstall bash-completion (or selectively remove stuff you don't
> - -bash: no job control in this shell
Did you log in on the hercules console itself? That is an extremely
limited 'dumb' console. You're better of using SSH.
> - C locale (no non-ASCII characters). Not sure why.
That's mainly because of the dumb console. You can generate other locales
if you prefer.
> 17. Something strange happened when I suspended reportbug: it seems
> to have accepted a ^C that was meant to be directed at bash.
No idea. Possibly related to the dumb console?
Not an installer issue though. If you want to follow up on it, please try
to narrow it down and file a separate report against the appropriate
> 18. Next step is to install telnet as an inetd service to get rid of
> some overhead.
Up to you :-)
> So, very smooth. I’m impressed.
As the installation was essentially successful and without issues, I'm
closing your report.
--- End Message ---