Thierry Parmentelat [Tue, 25 Feb 2020 10:04:47 +0000 (11:04 +0100)]
more debugging; show contents of /boot at intermediate stages
Thierry Parmentelat [Tue, 25 Feb 2020 10:07:58 +0000 (11:07 +0100)]
notify when creating site_admin account
Thierry Parmentelat [Fri, 29 Nov 2019 14:09:13 +0000 (15:09 +0100)]
turns out that [ -f ] returns true in bash !
Thierry Parmentelat [Fri, 29 Nov 2019 12:05:46 +0000 (13:05 +0100)]
no real change, but be a little more verbose when dealing with kernel
Thierry Parmentelat [Thu, 28 Nov 2019 14:40:37 +0000 (15:40 +0100)]
use python3 in helper script embedded in bash
Thierry Parmentelat [Mon, 7 Jan 2019 10:33:33 +0000 (11:33 +0100)]
Setting tag bootcd-6.0-0
requires 'genisoimage' rather than deprecated 'mkisofs'
set kernel arg `net.ifnames=0` in addition to `biosdevname=0`
does not use ifconfig, instead use ip and nm-ifup, or look up /sys/class/net
slightly review the contents of forensics messages
parmentelat [Tue, 18 Dec 2018 09:49:00 +0000 (10:49 +0100)]
turn off set -x in pl_netinit
parmentelat [Tue, 18 Dec 2018 09:48:16 +0000 (10:48 +0100)]
tentative hybrid code for 27 and 29
parmentelat [Mon, 17 Dec 2018 20:26:18 +0000 (21:26 +0100)]
trailing whitespaces by atom
parmentelat [Mon, 17 Dec 2018 20:25:53 +0000 (21:25 +0100)]
add net.ifnames=0 to the kernel args for still having eth0 on f29
parmentelat [Mon, 17 Dec 2018 20:19:32 +0000 (21:19 +0100)]
stop running dmesg as part of forensics
parmentelat [Mon, 17 Dec 2018 20:15:00 +0000 (21:15 +0100)]
cosmetic
parmentelat [Mon, 17 Dec 2018 20:14:47 +0000 (21:14 +0100)]
have pl_netinit source pl_functions, for messages to show up as expected
parmentelat [Fri, 14 Dec 2018 13:31:35 +0000 (14:31 +0100)]
untabified
parmentelat [Fri, 14 Dec 2018 10:39:26 +0000 (11:39 +0100)]
more harmless space-related matters
parmentelat [Fri, 14 Dec 2018 10:36:16 +0000 (11:36 +0100)]
untabified all init scripts
parmentelat [Fri, 14 Dec 2018 10:32:00 +0000 (11:32 +0100)]
cleanup comments
parmentelat [Thu, 13 Dec 2018 16:22:37 +0000 (17:22 +0100)]
for firing up the network: no longer rely on ifconfig
use /sys/class/net for checking for a device
and /usr/libexec/nm-ifup to turn it on
parmentelat [Tue, 4 Dec 2018 14:53:58 +0000 (15:53 +0100)]
rename mkisofs into genisoimage
Thierry Parmentelat [Sun, 10 Jul 2016 14:34:54 +0000 (16:34 +0200)]
Setting tag bootcd-5.4-1
use 8.8.8.8 s a DNS server if nothing else works
more consistent and verbose forensics system
Thierry Parmentelat [Fri, 8 Jul 2016 12:12:07 +0000 (14:12 +0200)]
bump version number to 5.4
Thierry Parmentelat [Fri, 8 Jul 2016 12:11:57 +0000 (14:11 +0200)]
a few notes
Thierry Parmentelat [Fri, 8 Jul 2016 09:41:58 +0000 (11:41 +0200)]
when resolv.conf is empty, always add 8.8.8.8
on top of DNS1 and DNS2
Thierry Parmentelat [Thu, 7 Jul 2016 16:14:50 +0000 (18:14 +0200)]
reviewed to ensure maximal coverage of forensics traces
Thierry Parmentelat [Tue, 8 Dec 2015 10:17:51 +0000 (11:17 +0100)]
Setting tag bootcd-5.3-3
patch for f23 as of dec. 2015 where kernel and initrd show up in
a new location under /boot
Thierry Parmentelat [Mon, 7 Dec 2015 09:47:22 +0000 (10:47 +0100)]
f23: patch for locating kernel to embed on bootcd
something weird seems to happen with rpm kernel-core, so we have
a second strategy to locate kernel
Thierry Parmentelat [Mon, 7 Dec 2015 08:18:57 +0000 (09:18 +0100)]
add check for at least one kernel found
Thierry Parmentelat [Fri, 13 Nov 2015 13:28:08 +0000 (14:28 +0100)]
Setting tag bootcd-5.3-2
various fixes for resolv.conf
Thierry Parmentelat [Wed, 26 Aug 2015 12:13:43 +0000 (14:13 +0200)]
fiz patch : resolv.conf as elaborated by systemd in /run looks like a special file
we can only overwrite the genuine file in /etc/resolv.conf, not the symlink
Thierry Parmentelat [Wed, 26 Aug 2015 09:49:59 +0000 (11:49 +0200)]
a lame attempt to manually fix /etc/resolv.conf on f22
Thierry Parmentelat [Wed, 26 Aug 2015 09:35:35 +0000 (11:35 +0200)]
add inspection of resolv.conf after network startup - f22 test node has an empty resolv.conf w/ a static config
Thierry Parmentelat [Fri, 26 Jun 2015 05:57:50 +0000 (07:57 +0200)]
Setting tag bootcd-5.3-1
use TLSv1 instead of SSLv3
on systemd distros, do not try to replace the whole boot sequence
as older bootCDs did; pl_boot is now just one more service, not the single
service to run; this solves f21 bootCDs not starting up properly
Thierry Parmentelat [Tue, 23 Jun 2015 14:01:15 +0000 (16:01 +0200)]
backing off on the idea of an early ssh access that is more harmful than helpful
Thierry Parmentelat [Tue, 23 Jun 2015 13:01:30 +0000 (15:01 +0200)]
less noide
Thierry Parmentelat [Tue, 23 Jun 2015 11:21:59 +0000 (13:21 +0200)]
open early ssh access to myplc's debug key
Thierry Parmentelat [Tue, 23 Jun 2015 11:21:29 +0000 (13:21 +0200)]
troubleshooting early ssh access
Thierry Parmentelat [Sat, 20 Jun 2015 22:39:32 +0000 (00:39 +0200)]
bump bootcd major number to 5.3
Thierry Parmentelat [Sat, 20 Jun 2015 22:38:10 +0000 (00:38 +0200)]
pl_boot uses tlsv1 instead of sslv3 to retrieve bootmanager code
tlsv1 has been checked to be supported and enabled in f14
Thierry Parmentelat [Fri, 19 Jun 2015 14:45:31 +0000 (16:45 +0200)]
at attempt to fix f21 boot sequence on physical nodes
Thierry Parmentelat [Wed, 17 Jun 2015 08:16:46 +0000 (10:16 +0200)]
gather more data when node is stuck without network
Thierry Parmentelat [Tue, 16 Jun 2015 08:18:53 +0000 (10:18 +0200)]
show ifconfig's output while waiting for the ethernet device
Thierry Parmentelat [Tue, 16 Jun 2015 08:10:52 +0000 (10:10 +0200)]
increase network initialization timeout to 2 minutes to check conjecture regarding poor system deps
Thierry Parmentelat [Tue, 16 Jun 2015 08:10:20 +0000 (10:10 +0200)]
cosmetic changes in the display when network interface cannot be brought up
Thierry Parmentelat [Fri, 24 Apr 2015 13:09:46 +0000 (15:09 +0200)]
Setting tag bootcd-5.2-6
remove default kargs systemd.log_target=console console=ttyS0,115200
simpler bootcd layout (no memdisk nor isolinux-debug.bin)
keep track of actual kargs in log file for build.sh
Thierry Parmentelat [Fri, 24 Apr 2015 13:07:57 +0000 (15:07 +0200)]
display line with initial '*'
Thierry Parmentelat [Fri, 24 Apr 2015 10:12:53 +0000 (12:12 +0200)]
remove default console=ttyS0,115200 and systemd.log_target=console from default kargs
I suspect this is what was causing f21 bootCDs to remain silent
Thierry Parmentelat [Fri, 24 Apr 2015 10:11:21 +0000 (12:11 +0200)]
new target sync-build to push build.sh on a test plc
Thierry Parmentelat [Thu, 23 Apr 2015 15:50:24 +0000 (17:50 +0200)]
keep track of original kernel in kernel.from
Thierry Parmentelat [Fri, 3 Apr 2015 05:50:25 +0000 (07:50 +0200)]
Setting tag bootcd-5.2-5
changes for fedora 21 and its new release of syslinux
Thierry Parmentelat [Mon, 9 Mar 2015 09:46:13 +0000 (10:46 +0100)]
update micro doc on the sync targets
Thierry Parmentelat [Mon, 9 Mar 2015 09:22:05 +0000 (10:22 +0100)]
stupid me
Thierry Parmentelat [Mon, 9 Mar 2015 08:02:06 +0000 (09:02 +0100)]
previous commit works fine, so just a bit of cleanup
Thierry Parmentelat [Sun, 8 Mar 2015 20:34:05 +0000 (21:34 +0100)]
fixed the logic for the timeout (break instead of continue..)
Thierry Parmentelat [Sun, 8 Mar 2015 20:16:27 +0000 (21:16 +0100)]
allow for a 30s delay before eth0 comes up
Thierry Parmentelat [Sun, 8 Mar 2015 18:39:16 +0000 (19:39 +0100)]
looks like our problem is that eth0 is not yet known to the system when we try to do the ifconfig
Thierry Parmentelat [Sun, 8 Mar 2015 18:38:28 +0000 (19:38 +0100)]
a convenience target to push initscripts/ and systemd/ in an unwrapped area
Thierry Parmentelat [Sun, 8 Mar 2015 17:04:52 +0000 (18:04 +0100)]
coming back to a model that works .. most of the time
Thierry Parmentelat [Sun, 8 Mar 2015 13:53:27 +0000 (14:53 +0100)]
mmmh
Thierry Parmentelat [Sat, 7 Mar 2015 21:43:14 +0000 (22:43 +0100)]
obviously I did not get it what ExecStart supports - just like anything else in systemd apparently
Thierry Parmentelat [Sat, 7 Mar 2015 14:54:09 +0000 (15:54 +0100)]
redirect pl_sysinit and pl_boot to /dev/console
Thierry Parmentelat [Fri, 6 Mar 2015 23:20:57 +0000 (00:20 +0100)]
attempts to write stuff on the wemu output
Thierry Parmentelat [Fri, 6 Mar 2015 23:17:10 +0000 (00:17 +0100)]
gettty on tty2 should be spawned as per an entry in pl_boot.target.wants/
Thierry Parmentelat [Fri, 6 Mar 2015 23:15:15 +0000 (00:15 +0100)]
see if we could get rid of pl_wrapper
Thierry Parmentelat [Fri, 6 Mar 2015 23:14:56 +0000 (00:14 +0100)]
this sounds better although it does not fix anything
Thierry Parmentelat [Fri, 6 Mar 2015 23:11:32 +0000 (00:11 +0100)]
fixes in the workflow to debug ping-node issues
Thierry Parmentelat [Thu, 5 Mar 2015 18:42:05 +0000 (19:42 +0100)]
implement some make targets to ease testing of a bootCD layout, esp. wrt the ping_node step
Thierry Parmentelat [Wed, 4 Mar 2015 13:11:18 +0000 (14:11 +0100)]
trying to add an explicit dependency to network-online.target
Thierry Parmentelat [Tue, 3 Mar 2015 15:02:29 +0000 (16:02 +0100)]
the kernel flag systemd.log_level=debug is not set by default anymore, callers need to pass the 'systemd-debug' option to GetBootMedium (or to set that tag on node)
Thierry Parmentelat [Mon, 2 Mar 2015 16:29:50 +0000 (17:29 +0100)]
reinstate systemd debug in bootcd phase, as we do not get to boot off the bootCD very reliably
Thierry Parmentelat [Sat, 28 Feb 2015 11:13:58 +0000 (12:13 +0100)]
reduce amount of messages issued by the bootcd
Thierry Parmentelat [Fri, 27 Feb 2015 13:58:36 +0000 (14:58 +0100)]
just use the shell 'function' syntax
Thierry Parmentelat [Fri, 27 Feb 2015 13:58:04 +0000 (14:58 +0100)]
tweaks for f21
Thierry Parmentelat [Fri, 30 Jan 2015 16:48:35 +0000 (17:48 +0100)]
no real change in terms of the function
Thierry Parmentelat [Tue, 27 Jan 2015 20:53:31 +0000 (21:53 +0100)]
temp version that at least fixes a serious regression in previous revision
Thierry Parmentelat [Tue, 27 Jan 2015 09:37:46 +0000 (10:37 +0100)]
fedora21 comes with syslinux v6 which requires ldlinux.c32 in addition to its new isolinux.bin
in the process, drop support for an old location of isolinux.bin (pre-f12)
Thierry Parmentelat [Fri, 21 Mar 2014 15:01:52 +0000 (16:01 +0100)]
Setting tag bootcd-5.2-4
no change, only some internal comments were changed
Thierry Parmentelat [Thu, 6 Feb 2014 10:12:11 +0000 (11:12 +0100)]
just correct a comment
Thierry Parmentelat [Wed, 26 Jun 2013 07:03:56 +0000 (09:03 +0200)]
Setting tag bootcd-5.2-3
attempt to make node take off more reliable with pl_wrapper
Thierry Parmentelat [Wed, 26 Jun 2013 06:23:07 +0000 (08:23 +0200)]
undo previous move
Thierry Parmentelat [Wed, 26 Jun 2013 05:35:58 +0000 (07:35 +0200)]
daemonize pl_wrapper to fulfil systemd
Thierry Parmentelat [Tue, 25 Jun 2013 16:29:45 +0000 (18:29 +0200)]
pl_wrapper needs chmod +x
Thierry Parmentelat [Tue, 25 Jun 2013 15:49:03 +0000 (17:49 +0200)]
attempt to gather more data on bootcd boot sequence when run under systemd
Thierry Parmentelat [Wed, 24 Apr 2013 13:31:16 +0000 (15:31 +0200)]
Setting tag bootcd-5.2-2
fix version number as it shows up in version.txt (was empty)
Thierry Parmentelat [Wed, 24 Apr 2013 06:50:57 +0000 (08:50 +0200)]
pass only 2 numbers to version.txt as bootmanager's logic depends on that
Thierry Parmentelat [Tue, 23 Apr 2013 09:59:40 +0000 (11:59 +0200)]
pass rpmversion to prep.sh instead of trying to recompute it from the specfile
in fact the specfile that prep.sh sees is the one that comes from git, which does not have
the header prepended, so some variables are not defined and rpm squeaks
Thierry Parmentelat [Mon, 22 Apr 2013 13:06:53 +0000 (15:06 +0200)]
attempt to fix build/version.txt in bootcd image
Thierry Parmentelat [Thu, 7 Mar 2013 07:29:16 +0000 (08:29 +0100)]
Setting tag bootcd-5.2-1
turn off systemd traces at boot-time
Thierry Parmentelat [Thu, 7 Mar 2013 07:13:01 +0000 (08:13 +0100)]
bump to 5.2
Thierry Parmentelat [Wed, 27 Feb 2013 17:24:06 +0000 (18:24 +0100)]
oops, really turn off systemd debugging
Thierry Parmentelat [Tue, 26 Feb 2013 11:48:10 +0000 (12:48 +0100)]
Setting tag bootcd-5.1-5
fix console on tty2: this broke the build on fedora14
even though f14 doesn't use systemd by defaults there is /etc/systemd in there
Thierry Parmentelat [Sat, 23 Feb 2013 16:08:06 +0000 (17:08 +0100)]
mkdir default.target.wants if not present, for f14
Thierry Parmentelat [Fri, 22 Feb 2013 17:57:37 +0000 (18:57 +0100)]
Setting tag bootcd-5.1-4
turn on getty/console on tty2
Stephen Soltesz [Thu, 21 Feb 2013 16:43:11 +0000 (08:43 -0800)]
enable getty@tty2 in parallel with pl_boot.service
Thierry Parmentelat [Thu, 21 Feb 2013 11:11:04 +0000 (12:11 +0100)]
Setting tag bootcd-5.1-3
reviewed for systemd & f18
Thierry Parmentelat [Wed, 20 Feb 2013 10:51:14 +0000 (11:51 +0100)]
turn off systemd logging on console
Thierry Parmentelat [Thu, 14 Feb 2013 11:07:53 +0000 (12:07 +0100)]
full messages to get on console when rebooting if DEBUG_SYSTEMD is set
Thierry Parmentelat [Wed, 13 Feb 2013 20:42:56 +0000 (21:42 +0100)]
redirect systemd log on console
Thierry Parmentelat [Sat, 9 Feb 2013 07:03:35 +0000 (08:03 +0100)]
cannot use kargs to select bootCD default systemd target - this would
apply to hdd boots as well; symlink default.target instead
Thierry Parmentelat [Fri, 8 Feb 2013 18:42:37 +0000 (19:42 +0100)]
start with the simplest possible setup
Thierry Parmentelat [Mon, 7 May 2012 10:28:30 +0000 (12:28 +0200)]
Setting tag bootcd-5.1-2
turn off selinux, turns out needed with some combinations like lxc/f14