Anaconda-devel December 2008

anaconda-devel@lists.fedoraproject.org

25 participants
100 discussions

Re: [PATCH] Append only one console argument in the yaboot.conf. (#235826)
by Joel Granados 15 Dec '08

15 Dec '08

----- "Hans de Goede" <hdegoede(a)redhat.com> wrote: > Joel Granados Moreno wrote: > > --- > > bootloaderInfo.py | 1 - > > 1 files changed, 0 insertions(+), 1 deletions(-) > > > > diff --git a/bootloaderInfo.py b/bootloaderInfo.py > > index 57caaf9..9829332 100644 > > --- a/bootloaderInfo.py > > +++ b/bootloaderInfo.py > > @@ -501,7 +501,6 @@ class bootloaderInfo: > > # don't keep the comma in the options > > self.serialOptions = options[1:] > > > > - self.args.append("console=tty0") > > self.args.append("console=%s%s" %(self.serialDevice, > options)) > > > > self.serial = 1 > Well, I spoke to my nearest kernel hacker and he said that the behavior with two arguments was undefined for rhel4 kernel. Moreover console seems to end up misconfigured after install. > Hmm, I'm not terrible famliar with this code but why do we want to > only have > one console argument, according to the kernel docs it is perfectly > fine to have can you point me to the docs, maybe this has some other cause. Maybe its user negligence, I see now that the second console argument was ttyS0 which is the default. Maybe what happened was that anaconda was installed without the console argument. Which is strange becuase that would mean that at installation time the person would have had to had access to the machine through other means. And I understand by report that that was most unlikely. If you are correct this is most likely not a bug. :) > multiple console=foo options, any kernel output will be send to all of > them, And will receive input from all of them as well. right? > and a process opening /dev/console will get the last console= > argument. So I do > not understand what exactly this is supposed to fix ? > > Regards, > > Hans Well, I spoke to my nearest kernel hacker and he said that the behavior with two arguments was undefined for rhel4 kernel, can you point me to the docs, maybe this has some other cause. Moreover console seems to end up misconfigured after install > > _______________________________________________ > Anaconda-devel-list mailing list > Anaconda-devel-list(a)redhat.com > https://www.redhat.com/mailman/listinfo/anaconda-devel-list -- Joel Andres Granados Red Hat / Brno Czech Republic

2 1

[PATCH] booty. make sure we write only one console arg.
by Joel Granados Moreno 15 Dec '08

15 Dec '08

hey list: This is a patch for RHEL4 booty. We are adding two console kernel parameters when we really should make sure to pass only one. This patch appends whatever was passed to the console argument at install time. If no value was passed then we default to ttyS0 with not options. Review greatly appreciated.

2 2

PATCH[RHEL4]: mdraid1-allow-bootloader-on-mbr
by Hans de Goede 15 Dec '08

15 Dec '08

Hi All, Basicly the same patch I've just committed for master. This patch allows installing the bootloader on the mbr when using mdraid1, as requested in bug 217176. It also contains some special handling of this case in bootloader.py, to make sure our default behaviour does not change. Regards, Hans diff --git a/bootloader.py b/bootloader.py index 486cb64..d88da31 100644 --- a/bootloader.py +++ b/bootloader.py @@ -85,7 +85,8 @@ def bootloaderSetupChoices(dispatch, bl, fsset, diskSet, dir): log("MBR not suitable as boot device; installing to partition") bl.defaultDevice = "boot" bl.setDevice(choices[bl.defaultDevice][0]) - elif choices and choices.has_key("mbr"): + elif choices and choices.has_key("mbr") and not \ + (choices.has_key("boot") and choices["boot"][1] == N_("RAID Device")): bl.setDevice(choices["mbr"][0]) elif choices and choices.has_key("boot"): bl.setDevice(choices["boot"][0]) diff --git a/fsset.py b/fsset.py index 22fd824..cee9ef7 100644 --- a/fsset.py +++ b/fsset.py @@ -1340,6 +1340,11 @@ MAILADDR root if bootDev.getName() == "RAIDDevice": ret['boot'] = (bootDev.device, N_("RAID Device")) + try: + # we won't have this on zFCP-only zSeries systems + ret['mbr'] = (bl.drivelist[0], N_("Master Boot Record (MBR)")) + except: + pass return ret if iutil.getPPCMacGen() == "NewWorld":

2 1

PATCH[RHEL53]: fix traceback on ibft systems when not using chap
by Hans de Goede 12 Dec '08

12 Dec '08

Hi All, The subject says it all, this should fix 432819 Regards, Hans diff --git a/iscsi.py b/iscsi.py index 37366cb..89220c0 100644 --- a/iscsi.py +++ b/iscsi.py @@ -388,10 +388,20 @@ class iscsi(object): self.updateIscsidConf(record) + if record.has_key('node.session.auth.username'): + username = record['node.session.auth.username'] + else: + username = None + + if record.has_key('node.session.auth.password'): + password = record['node.session.auth.password'] + else: + password = None + t = iscsiTarget(ipaddr=record['node.conn[%s].address' % (n,)], port=record['node.conn[%s].port' % (n,)], - user=record['node.session.auth.username'], - pw=record['node.session.auth.password'], + user=username, + pw=password, discover = False) self.targets.append(t)

2 1

Include modules so isw raid can work in RHEL-5
by Joel Granados Moreno 12 Dec '08

12 Dec '08

In my test box the isw raid doesn't even get recognized because of missing modules in the installer environment. Not sure If I have the complete list of modules. Review greatly appreciated.

2 2

[PATCH] 474937 - don't duplicate --dport 22
by Chris Lumens 12 Dec '08

12 Dec '08

We were writing --dport 22 to /etc/sysconfig/iptables twice beacuse ssh was getting included both in the kickstart file and implicitly in firewall.py. Check for that duplication. While I'm at it, also actually handle --service= in the kickstart file in firewall.py. - Chris diff --git a/firewall.py b/firewall.py index f3e689d..ef543e6 100644 --- a/firewall.py +++ b/firewall.py @@ -35,7 +35,7 @@ class Firewall: self.enabled = 1 self.trustdevs = [] self.portlist = [] - self.servicelist = ["ssh"] + self.servicelist = [] def writeKS(self, f): f.write("firewall") @@ -54,7 +54,10 @@ class Firewall: if not self.enabled: args.append("--disabled") return args - + + if not "ssh" in self.servicelist and not "22:tcp" in self.portlist: + args += ["--service=ssh"] + for dev in self.trustdevs: args = args + [ "--trust=%s" %(dev,) ] diff --git a/kickstart.py b/kickstart.py index 4d70d51..961da30 100644 --- a/kickstart.py +++ b/kickstart.py @@ -247,6 +247,9 @@ class Firewall(commands.firewall.F10_Firewall): for port in self.ports: self.handler.id.firewall.portlist.append (port) + for svc in self.services: + self.handler.id.firewall.servicelist.append (svc) + class Firstboot(commands.firstboot.FC3_Firstboot): def parse(self, args): commands.firstboot.FC3_Firstboot.parse(self, args)

3 2

PATCH: allow installing grub on the MBR if /boot is on mdraid.
by Hans de Goede 11 Dec '08

11 Dec '08

Hi All, This patch removes a special case when /boot (or / without a separate /boot) is on mdraid. For some reason we only allow installing grub to the mdX device then instead of allowing both the mbr of the 1st disk and the device holding /boot. I've searched through the anaconda history and this special handling of mdraid /boot has been there since before Feb. 2002, it was inherited from before booty was introduced and it does not seem to make sense. Either we cannot have /boot on mdraid at all (in all cases except for raid1) or we can have /boot on dmraid (in case of raid1) and then it does not matter if the bootstrap of grub starts on the mbr or on the first sector of a partition. This fixes 217176, note that is a RHEL4 bug but this patch is against master as it still happens there too. Next is fixing this for RHEL5 and then RHEL4. Regards, Hans diff --git a/fsset.py b/fsset.py index c9cd098..b726985 100644 --- a/fsset.py +++ b/fsset.py @@ -1522,10 +1522,6 @@ MAILADDR root ret['boot'] = (bootDev.device, N_("EFI System Partition")) return ret - if bootDev.getName() == "RAIDDevice": - ret['boot'] = (bootDev.device, N_("RAID Device")) - return ret - if iutil.getPPCMacGen() == "NewWorld": ret['boot'] = (bootDev.device, N_("Apple Bootstrap")) n = 1

4 11

PATCH: fix various *syntax* errors in master
by Hans de Goede 11 Dec '08

11 Dec '08

Hi all, This patch fixes the following syntax errors caught by pychecker: iw/lvm_dialog_gui.py:1024: Invalid arguments to (createSuggestedVGName), got 1, expected 2 iw/welcome_gui.py:67: No global (sys) found textw/network_text.py:75: No global (herrors) found textw/welcome_text.py:44: No global (sys) found I choose to rename neterrors to herrors, instead of replacing the use of the undefined herrors with neterrors, to make things identical between textmode and gui mode, to avoid copy and paste errors like this one in the future. Regards, Hans diff --git a/iw/lvm_dialog_gui.py b/iw/lvm_dialog_gui.py index 58793b7..25c0e6d 100644 --- a/iw/lvm_dialog_gui.py +++ b/iw/lvm_dialog_gui.py @@ -967,7 +967,7 @@ class VolumeGroupEditor: self.dialog.destroy() self.dialog = None - def __init__(self, partitions, diskset, intf, parent, origvgrequest, isNew = 0): + def __init__(self, anaconda, partitions, diskset, intf, parent, origvgrequest, isNew = 0): self.partitions = partitions self.diskset = diskset self.origvgrequest = origvgrequest @@ -1021,7 +1021,7 @@ class VolumeGroupEditor: if not self.isNew: self.volnameEntry.set_text(self.origvgrequest.volumeGroupName) else: - self.volnameEntry.set_text(lvm.createSuggestedVGName(self.partitions)) + self.volnameEntry.set_text(lvm.createSuggestedVGName(self.partitions, anaconda.id.network)) else: lbl = createAlignedLabel(_("Volume Group Name:")) self.volnameEntry = gtk.Label(self.origvgrequest.volumeGroupName) diff --git a/iw/partition_gui.py b/iw/partition_gui.py index 384758d..b2c11fd 100644 --- a/iw/partition_gui.py +++ b/iw/partition_gui.py @@ -1146,7 +1146,8 @@ class PartitionWindow(InstallWindow): return 1 def editLVMVolumeGroup(self, origvgrequest, isNew = 0): - vgeditor = lvm_dialog_gui.VolumeGroupEditor(self.partitions, + vgeditor = lvm_dialog_gui.VolumeGroupEditor(self.anaconda, + self.partitions, self.diskset, self.intf, self.parent, origvgrequest, isNew) diff --git a/iw/welcome_gui.py b/iw/welcome_gui.py index 8dc1a7e..d2e6fc5 100644 --- a/iw/welcome_gui.py +++ b/iw/welcome_gui.py @@ -19,6 +19,7 @@ import gtk import gui +import sys from iw_gui import * from constants import * diff --git a/textw/network_text.py b/textw/network_text.py index 0b90405..f35510f 100644 --- a/textw/network_text.py +++ b/textw/network_text.py @@ -59,7 +59,7 @@ class HostnameWindow: return INSTALL_BACK hostname = string.strip(hostEntry.value()) - neterrors = network.sanityCheckHostname(hostname) + herrors = network.sanityCheckHostname(hostname) if not hostname: ButtonChoiceWindow(_("Error with Hostname"), @@ -68,7 +68,7 @@ class HostnameWindow: buttons = [ _("OK") ]) continue - if neterrors is not None: + if herrors is not None: ButtonChoiceWindow(_("Error with Hostname"), _("The hostname \"%s\" is not valid for the " "following reason:\n\n%s") diff --git a/textw/welcome_text.py b/textw/welcome_text.py index c0f771d..bf58b58 100644 --- a/textw/welcome_text.py +++ b/textw/welcome_text.py @@ -20,6 +20,7 @@ from snack import * from constants_text import * from constants import * +import sys import gettext _ = lambda x: gettext.ldgettext("anaconda", x)

2 1

Doing a reboot in %post of ks.cfg
by Niels de Vos 11 Dec '08

11 Dec '08

Hello, with anaconda-10.2.x we did a reboot in %post with 'kill -SIGINT 1'. The loader catches the signal an does a reboot. With anaconda-11.1.x this still functions, but gives an ugly stacktrace (see anacdump-stack.txt attached) in the main window (at least text-mode). What is the advised method of automatic reboot in %post? If this is a not-implemented-yet feature, what would be a good solution for a non-interactive install+reboot? Thanks, Niels Traceback (most recent call first): File "/tmp/treedir.20081/instimage/usr/lib/anaconda/iutil.py", line 50, in execWithRedirect raise RuntimeError, "Error running " + command + ": " + msg File "/usr/lib/anaconda/kickstart.py", line 64, in run root = scriptRoot) File "/usr/lib/anaconda/kickstart.py", line 783, in <lambda> map (lambda s: s.run(anaconda.rootPath, serial, anaconda.intf), postScripts) File "/usr/lib/anaconda/kickstart.py", line 783, in postAction map (lambda s: s.run(anaconda.rootPath, serial, anaconda.intf), postScripts) File "/tmp/treedir.20081/instimage/usr/lib/anaconda/packages.py", line 44, in doPostAction anaconda.id.instClass.postAction(anaconda, flags.serial) File "/usr/lib/anaconda/dispatch.py", line 201, in moveStep rc = stepFunc(self.anaconda) File "/usr/lib/anaconda/dispatch.py", line 124, in gotoNext self.moveStep() File "/usr/lib/anaconda/text.py", line 588, in run anaconda.dispatch.gotoNext() File "/usr/bin/anaconda", line 982, in ? anaconda.intf.run(anaconda) RuntimeError: Error running /bin/sh: Interrupted system call Local variables in innermost frame: chroot: <function chroot at 0xadf5aac> stdout: <open file '/dev/tty3', mode 'w' at 0xb16b458> errno: 4 stdin: <open file '/dev/tty3', mode 'r' at 0xb16b4a0> argv: ['/tmp/ks-script-xY-HcO'] searchPath: 0 command: /bin/sh stderr: <open file '/dev/tty3', mode 'w' at 0xb16b410> msg: Interrupted system call root: /mnt/sysimage proc: <subprocess.Popen object at 0xaafb16c>

2 2

PATCH: install the bootloader on partition instead of mbr by default in raid1 case
by Hans de Goede 10 Dec '08

10 Dec '08

Hi all, As previously discussed we do want to allow installing the bootloader on the MBR in the raid 1 case (which until some time ago we did not allow) (#217176), but we do not want to change our current default of installing to the partition (which gives is the mirror redundancy, due to special handling in booty). This patch effectively reverts commit 48d582347492493c87b2b368394fd8e7342688bc, which removed the raid1 case special handling, and then adds mbr support to the raid special case in fsset.py, and add some special case handling to bootloader.py to not default to the mbr in this case. Regards, Hans diff --git a/bootloader.py b/bootloader.py index 29bb9c2..dcc9899 100644 --- a/bootloader.py +++ b/bootloader.py @@ -119,7 +119,8 @@ def bootloaderSetupChoices(anaconda): anaconda.id.bootloader.setDevice(choices[anaconda.id.bootloader.defaultDevice][0]) elif choices and iutil.isMactel() and choices.has_key("boot"): # haccckkkk anaconda.id.bootloader.setDevice(choices["boot"][0]) - elif choices and choices.has_key("mbr"): + elif choices and choices.has_key("mbr") and not \ + (choices.has_key("boot") and choices["boot"][1] == N_("RAID Device")): anaconda.id.bootloader.setDevice(choices["mbr"][0]) elif choices and choices.has_key("boot"): anaconda.id.bootloader.setDevice(choices["boot"][0]) diff --git a/fsset.py b/fsset.py index 9d798eb..497ccfa 100644 --- a/fsset.py +++ b/fsset.py @@ -1537,6 +1537,11 @@ MAILADDR root ret['boot'] = (bootDev.device, N_("EFI System Partition")) return ret + if bootDev.getName() == "RAIDDevice": + ret['boot'] = (bootDev.device, N_("RAID Device")) + ret['mbr'] = (bl.drivelist[0], N_("Master Boot Record (MBR)")) + return ret + if iutil.getPPCMacGen() == "NewWorld": ret['boot'] = (bootDev.device, N_("Apple Bootstrap")) n = 1

2 1

← Newer
1
2
3
4
5
6
7
8
9
10
Older →

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Anaconda-devel December 2008