Waiki Wright píše v Út 06. 09. 2011 v 18:27 -0400:
> Hello Dan,
> I saw your name in one of the bugs related to device_cio_free.conf and
> hope that you may be able to answer my question. If you prefer me to
> send my inquire to a mailing list, please let me know.
I'm sending the reply to mailing list, so more people can comment.
> I am using s390utils-1.8.2-28.el6.s390x package and I am wondering if
> there will be a race condition between rcS and device_cio_free.
> The device_cio_free.conf file said that it will start when rcS.conf is
> # free all devices on startup
> start on starting rcS
> console output
> exec /sbin/device_cio_free
> However, in the /sbin/device_cio_free script, it depends on /proc is
> mounted before the script is called:
> if [ ! -f $BLACKLIST ]; then
> echo "Error: $BLACKLIST kernel interface doesn't exist"
> exit 2
> If the file system is mounted later (from rcS), then device_cio_free
> script will fail. So should it be better if
> device_cio_free.conf to start on after /proc is mounted?
> Your thought?
Your conclusion looks logical. However this upstart job was proposed by
our upstart maintainer when I asked him for a solution to the following
problem: need to start an action before the system will try to mount
non-root filesystem devices or bring network interfaces up. Also it
shouldn't interfere with the startup process on primary arches.
Staring with Fedora 15 there is a systemd solution, see
device_cio_free.service in s390utils-1.14.0-1.fc16 
> Phil P wrote:
> So I pulled the latest Source code snapshot - 7747 compiled it - and tried
to boot the install images -
it just loops never getting past - the Linux message - "Trying to unpack
rootfs image as initramfs..."
> Went back to 7702 and it came up to the network configuration fine.
Compiled 7702 and latest 7747 today and both give similar result of 23 minutes or so and 33 Billion instruction count before unpack is finished.
14:15:10 Ö 15.455934! Trying to unpack rootfs image as initramfs...
14:38:09 Ö 704.650414! Freeing initrd memory: 77722k freed
> Are you running hercules under Windows or Linux?
> It also depends if you are using CTCs - it would be the IP address at the
other end, if you are using LCS devices it should be your router or gateway
to the internet. LCS Devices are simpler to use on Linux you do not need
to add routes just a change to sysctl.conf to
net.ipv4.conf.default.proxy_arp = 1
Keeping the configuration which worked for FC11 which used CTC.
If I don't run the hercules.init script before hercules startup the 2nd CTC fails to initalize due to tun (?) not available.
My host only has 1Gb RAM and 512Mb swap partition. I added a 2Gb swap file and with SVN 7747 don't seem to get segmentation faults anymore up IPL of C. Down in the interactive install it askes for 'partner to partner' and then states it is trying gateway and later DNS IP address. I haven't gotten the right info yet as it never cannot ping the DNS Server.
A little bit of my log:
14:38:36 Scanning for available network devices...
14:38:38 Autodetection found 0 devices.
14:38:40 Note: There is a device blacklist active× (Clearing might take long)
14:38:40 c) clear blacklist, m) manual config, r) rescan, s) shell:
15:13:19 HHC00160I SCP command: 'c'
15:13:19 Clearing device blacklist...
15:13:26 /init: line 491: nettable: bad array subscript
15:13:27 Network type (qeth, lcs, ctc, ? for help). Default is qeth:
15:13:29 INFO kernel:Ö 1764.690769! lcs: Loading LCS driver
15:13:31 INFO kernel:Ö 1765.561956! ctcm: CTCM driver initialized
15:13:34 HHC00160I SCP command: 'ctc'
15:13:34 Read,write channel (e.g. 0.0.0600,0.0.0601 or ? for help)
15:13:46 HHC00160I SCP command: '0.0.0600,0.0.0601'
15:13:48 Detected: channel-to-channel adapter (CTC/A)
15:13:48 CTC protocol (0, 1, 3, or ? for help). Default is 0:
15:13:54 HHC00160I SCP command: '0'
15:13:54 Activating network device...
15:13:55 INFO kernel:Ö 1777.926703! net ctc0: setup OK : r/w = ch-0.0.0600/ch-0.0.0601, p
15:13:55 rotocol : 0
15:13:58 Hostname of your new Linux guest (FQDN e.g. s390.redhat.com or ? for help):
15:13:58 INFO kernel:Ö 1779.297699! net ctc0: Connected with remote side
15:14:19 HHC00160I SCP command: 's390.halfmeg.org'
15:14:20 IPv4 address (e.g. 10.0.0.2 or ? for help)
15:15:05 HHC00160I SCP command: '192.168.200.3'
15:15:06 IPv4 address of your point-to-point partner or ? for help:
15:15:14 HHC00160I SCP command: '?'
15:15:14 Help text for point-to-point partner:
15:15:14 IPv4 address of your CTC or ESCON point-to-point partner.
15:15:14 IPv4 address of your point-to-point partner or ? for help:
15:15:46 HHC00160I SCP command: '192.168.200.4'
15:15:46 Trying to reach gateway 192.168.200.4...
15:15:53 IPv4 addresses of DNS servers (separated by colons ':' or ? for help):
15:16:01 HHC00160I SCP command: '?'
15:16:01 Help text for DNS servers:
15:16:01 Enter IPv4 addresses of DNS servers separated by colons ':'
15:16:01 Default are no DNS servers at all.
15:16:01 However, you might encounter problems without a nameserver,
15:16:01 especially with FTP installs.
15:16:01 An example with 2 servers would be: 10.0.0.250:10.1.1.1
15:16:01 IPv4 addresses of DNS servers (separated by colons ':' or ? for help):
15:18:12 HHC00160I SCP command: '220.127.116.11'
15:18:13 Trying to reach DNS servers...
15:19:14 Could not ping DNS server: 18.104.22.168
I tried 192.168.200.1 for point-to-point ( gateway ? ) which it didn't balk at but still couldn't get past the DNS address ( thinik I tried 192.168.1.254 for DNS also ).
You mentioned LCS is simpler so may attempt to switch to that.
Ignoring the above problems with DNS and perhaps the gateway addresses I continued the install and got to this point:
15:28:19 Starting sshd to allow login over the network.
15:28:19 HHC00801I Processor CP00: 'Operation exception' code 0001 ilc 4
15:28:19 HHC02267I PSW=0705F00180000000 000003FFFCF524A0 INST=B93E0002 KIMD 0,2 compute_intermediate_message_digest
15:28:19 HHC02267I V:0000000000000000: Translation exception 003B
15:28:19 HHC02267I V:000002AAD057EA60:K:06=67452301 EFCDAB89 98BADCFE 10325476 . Á.ï.¨iqª.ú...ä
15:28:19 /init: line 178: 364 Illegal instruction /usr/sbin/sshd -f /etc/ssh/sshd_c
15:28:19 Connect now to 192.168.200.3 and log in as user 'install' to start the installat
15:28:19 E.g. using: ssh -x install(a)192.168.200.3
15:28:19 For VNC or text mode, disable X11 forwarding (recommended) with 'ssh -x'.
15:28:19 For X11, enable X11 forwarding with 'ssh -X'.
15:28:19 You may log in as the root user to start an interactive shell.
15:29:14 HHC00160I SCP command: 'login root'
15:29:14 login root
15:29:14 sh: cannot set terminal process group (-1): Inappropriate ioctl for device
15:29:14 sh: no job control in this shell
15:29:26 HHC00160I SCP command: 'ls'
15:29:26 bin etc init lib64 proc run selinux tmp var
15:29:26 dev firmware lib modules root sbin sys usr
15:29:40 HHC00160I SCP command: 'shutdown -h now'b
A ssh from another terminal window on the same host pc timed out connecting. Logging in via the hercules console it does get to where one can do a 'ls' and get response back.
In a couple of places I noticed hercules get an 'operation exception' and the above snippet of the log shows one. Is this type error expected ?
> Phil Pinto wrote:
> Which version of Hercules are you using? I imagine if you are using the new
> FP stuff in hercules you are higher than SVN 7709
SVN 7731 - compiled 7747 this am but no testing done with it yetbbbbbbbbbbb
> I am using SVN 7704 with no issues running FC15 / yum or anything else..
I could not get SVN 7704 to compile and didn't see where any snapshot archive where it was successfully created either.
> I would shut-down hercules and make a backup copy of the Hercules DASD that
> you installed FC15 on before continuing so you do not have to go through the
> install again if things break later.
You may have misunderstood. I haven't created any Fedora DASD. I have been using the 'turnkey images' of FC11 & FC14 available at:
> This Hercules Disk image you created for FC15 hopefully has network
> connectivity through a CTC or LCS device and has access to the internet. If
> that is all working correctly.....
While messing with some of this lately, I have discovered the hercules.init script located at the above link causes hercules to get a segmentation fault when specific choices are made during IPL. It works fine with FC11, but with FC14 if one selects one of the two choices 0 or 1 instead of letting the default linux kernel boot, a segmentation fault occurs. If left to time out, the default kernel starts to boot and continues until logon prompt is displayed.
So as to leave most of old stuff behind I tried to boot FC15 using the kernel.img, initrd.img & generic.prm from
Using the hercules.init script mentioned above with minor changes for FC15 DASD and hercules version line, upon IPL of C, hercules gives a segmentation fault almost immediately. When I used an older version of hercules, or started hercules without the script ( don'r remember which one worked), I get to a point where it tells me a blacklist exists and do I want to clear it or continue. It also starts asking about a gateway IP address ??? Is that my Host OS leased IP address it grabbed via automagic means when I booted my host PC ?
Do I have to have the modules for TUN and brothers loaded ( via the hercules.init script ) in order to get access to anything from FC15 ?