summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
...
| | * mtu is a string from kickstart, not number (#832332)Jesse Keating2012-07-181-2/+2
| | |
| | * makeupdates: change to use install_to_dirBrian C. Lane2012-07-181-2/+2
| | |
| | * iscsi: fall back to session initiator name for null offload name (#836307)Radek Vykydal2012-07-121-0/+5
| | | | | | | | | | | | Resolves: rhbz#836307
| | * Accept iSCSI details on F12 press (#823447)Jesse Keating2012-07-091-1/+1
| | | | | | | | | | | | | | | | | | When F12 (the global default continue hotkey) is pressed the button value is None. But we still want to process the details when this happens. The use of bool here makes it a little easier to read.
| | * Check passphrase unless explicit cancel (#821724)Jesse Keating2012-07-091-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | snack/newt returns None for a button if F12 is pressed which is a global default. There can be passphrase data in the box when this happens and we should read it. The only time we should ignore anything in the box is if the user has explicitely pressed the cancel button. If the passphrase box is empty but the user pressed F12 we still handle that correctly as an implicit cancel.
| | * Reset selinux context on /etc/zipl.conf (#818875)Jesse Keating2012-07-091-1/+2
| | | | | | | | | | | | | | | | | | /etc/zipl.conf is written by anaconda when we install the boot loader. Because we mess with it, we have to reset the selinux context on it before rebooting.
| | * Let kickstart users log to device nodes (#835563)Brian C. Lane2012-07-091-9/+10
| | | | | | | | | | | | | | | | | | | | | | | | If a kickstart pre/post fails we read the logfile and add it to the anaconda log. This doesn't work well for device nodes like /dev/console so skip reading the log it isn't a regular file. Resolves: rhbz#835563
| | * Do not --force systemctl reboot (#818581)Radek Vykydal2012-07-091-4/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Forced reboot just kills processes without shutting units/services down. In case of NetworkManager.service this terminating of NM is racing with that of dhclient. If dhclient dies first, NM brings down respective connection which - if used by NFS mounted in dracut (e.g. for repo=nfs: repo=nfsiso: installs) - causes hang when systemd umounts filesystems. Resolves: rhbz#818581
| | * Make zipl call sensitive to non-zero exit (#730023)Jesse Keating2012-07-051-1/+2
| | | | | | | | | | | | | | | If zipl doesn't succeed, the install won't reboot when finished. This is not an acceptable situation.
| | * Provide option to make non-zero exits fatalJesse Keating2012-07-051-1/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | execWithCapture normally doesn't care about non-zero exits in called processes. In some cases (most?) this is preferred. In few cases we do actually care about the success or failure of the execution. "fatal" provides a boolean that will allow callers to decide if they care or not. An OSError is raised because there is already code to catch that in the except section. Related: rhbz#730023
| | * Keep format _mountType in sync with _type (#828650)Jesse Keating2012-07-021-0/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | Device formats have both a type and mountType property. The mountType property has a fallback of reading the _type attribute, if _mountType isn't set. But when we're migrating filesystems, the _mountType will have already been set, so just modifying _type won't effect any code looking at mountType. This can lead to the wrong data being written out to say /etc/fstab.
| | * Make use of common code to fetch IPJesse Keating2012-06-282-34/+14
| | |
| | * Add method to grab the first non-localhost IPJesse Keating2012-06-281-0/+21
| | | | | | | | | | | | This is mostly code that was in vnc.py as well as anaconda itself.
| | * Add readonly disks to the ignored list (#772977)Jesse Keating2012-06-281-0/+14
| | | | | | | | | | | | | | | If your disk is read-only, we're just going to flat out ignore it and not offer it to you for anything.
| | * Limit the maximum size of swap partitionsVratislav Podzimek2012-06-271-0/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | swapon creates a table using 1 B for each page in the swap. Therefore when a huge swap (e.g. 500 GB) is created, swapon fails with ENOMEM, because it is using vmalloc that allocates a continuous memory. Thus we need to limit the maximum size of swap partitions. And since we don't know the strict limit, let's use 128 GB that seems to work and should be enough for everybody. (ported 35679a70eab52b3b040a61fce9d0e43dbb977c98 from rhel6-branch)
| | * Remove swapAmount since it is not used anywhereVratislav Podzimek2012-06-271-17/+0
| | |
| | * Move swapSuggestion to storage and use a new suggested algorithm for itVratislav Podzimek2012-06-274-50/+69
| | | | | | | | | | | | | | | | | | | | | | | | | | | We were using an algorithm that suggested the swap size as 2 GB + size of RAM, but this resulted in huge swaps on machines with a lot of RAM. The new algorithm comes from the discussion with other teams. (ported 84b3444a277b73abeaddf7d4b186a79569eb56d2 from rhel6-branch) (ported 37415063594d00c896ff3c50496582f0c4e9e3d9 from rhel6-branch)
| | * Add missing '-fcoe' suffix to fipvlan -f option (#823086)Radek Vykydal2012-06-261-2/+4
| | | | | | | | | | | | Related: rhbz#823086
| | * Adapt to fipvlan naming change (#823086)Radek Vykydal2012-06-261-2/+2
| | | | | | | | | | | | | | | | | | The "-f" option keeps former naming with "-fcoe" suffix. Resolves: rhbz#823086
| | * Find initiator name for partial offload iscsi bnx2i devices (#830756)Radek Vykydal2012-06-261-3/+5
| | | | | | | | | | | | Resolves: rhbz#830756
| | * Handle offload qla4xxx iscsi devices properly (#823810)Radek Vykydal2012-06-263-21/+70
| | | | | | | | | | | | | | | | | | | | | | | | These kind of devices, when qla4xxx.ql4xdisablesysfsboot=1, is discovered by anaconda/udev as iscsi device yet it can not be handled by iscsiadm so we need to treat it specially. Resolves: rhbz#823810
| | * Set network configuration for root on partial offload iscsi devices (#819132)Radek Vykydal2012-06-261-2/+2
| | | | | | | | | | | | | | | | | | Set NM_CONTROLLED to "no" and pass ip= and ifname= parameters to dracut. Resolves: rhbz#819132
| | * makeupdates: add rpm contents to updates.imgBrian C. Lane2012-06-221-2/+15
| | | | | | | | | | | | | | | | | | This adds the -a option to makeupdates, it will expand the contents of an rpm into the root of the updates image. This overlays the installer's root filesystem, allowing any file to be updated.
| | * Fix handling of invalid bios disks (#819721)Jesse Keating2012-06-211-1/+1
| | |
| | * Use the correct path to kickstart files (#833070)Jesse Keating2012-06-211-1/+1
| | |
| | * We should use /usr/bin/pythonMartin Sivak2012-06-211-23/+23
| | | | | | | | | | | | (cherry picked from commit 3f1abc384736f0e105b46612e0601a50784a9528)
| | * Actually start anaconda-cleanup-initramfs.serviceWill Woods2012-06-211-0/+1
| | | | | | | | | | | | | | | | | | | | | Rule #1 of Computing: It Works Better When You Turn It On. Make anaconda.target request anaconda-cleanup-initramfs.service so it actually gets started.
| | * spec fix: include anaconda-cleanup-initramfsWill Woods2012-06-201-0/+1
| | | | | | | | | | | | Forgot to add it to the specfile, so builds failed. Whoops.
| * | Do not run Gtk.main() from a separate thread in exception handlingVratislav Podzimek2012-08-031-2/+25
| | | | | | | | | | | | | | | | | | Multiple threads drawing in a Gtk.main loop cause Gtk crash. To prevent such crashes, we need to run exception handling dialogs in the same thread as the main loop (if running).
| * | Do not run __del__() on anaconda interface in runDebugVratislav Podzimek2012-08-031-2/+0
| | | | | | | | | | | | UserInterface object no longer has the __del__() method.
| * | Add mainExceptionWindow and saveExceptionWindow methods to the interfacesVratislav Podzimek2012-08-032-0/+23
| | | | | | | | | | | | | | | | | | | | | python-meh has its own methods, but we have our own exception handling code (inherited from the python-meh's one) that needs access to both python-meh UI and anaconda UI. Hence anaconda's UIs need a mainExceptionWindow and saveExceptionWindow methods calling the python-meh's ones (GUI or TUI).
| * | Update the anaconda.pot file for our new translatable strings.Chris Lumens2012-08-021-3452/+1071
| | |
| * | Select language-specific groups as part of doInstall.Chris Lumens2012-08-023-5/+21
| | | | | | | | | | | | | | | This means you get the whatever-support group installed if you have the whatever language selected during installation.
| * | Correct the location of the detailederror.glade file.Chris Lumens2012-08-021-1/+1
| | |
| * | Extended partitions containing in-use logical partitions are also in-use.Chris Lumens2012-08-021-0/+4
| | |
| * | Add packaging log and scriptlet loggingBrian C. Lane2012-08-024-4/+33
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This adds a new logfile for packaging (/tmp/packaging.log) which will contain everything related to yum and rpm as well as other package systems when they are added. This also re-adds logging of the rpm scriptlet output. The 'Installing ...' messages end up being logged twice. Because rpm only logs to a file we capture the scriptlet output along with the package install message so that you can associate scriptlet problems with their package. This adds this file's log to our logs after the transaction is complete (or fails).
| * | Remove backend_logBrian C. Lane2012-08-024-136/+2
| | | | | | | | | | | | | | | | | | The second instance of rsyslog is not needed, and it interferes with virtio logging. The yum and rpm file io logging will be dealt with in another patch.
| * | Start rescue mode from the right place in the anaconda script.Chris Lumens2012-08-011-118/+121
| | | | | | | | | | | | | | | This patch looks scarier than it is, laregely because it requires unindenting a bunch of code to not have a rescue mode test involved.
| * | Update rescue mode to continue working in the newui world.Chris Lumens2012-08-012-90/+61
| | | | | | | | | | | | | | | | | | | | | | | | This is still just the existing newt rescue mode interface, but with adding advanced devices removed (requires other code that no longer exists). Of course, this all needs to be removed and replaced by something that does not use newt in the future but for now, it's enough of a stop-gap to fulfill release requirements.
| * | The text network config screen should take an intf, not an anaconda.Chris Lumens2012-08-011-7/+6
| | | | | | | | | | | | | | | | | | This is needed to have it work in the rescue mode interface, though it has not yet been tested due to the difficulty of testing an updates image without active networking.
| * | Remove upgrade.py and associated tests.Chris Lumens2012-08-013-583/+0
| | | | | | | | | | | | | | | All upgrading will be handled by an external process, and this gets rid of callers to some storage functions that I need to make rescue mode work again.
| * | Move the contents of bindMountDevDirectory to storage, where it's called.Chris Lumens2012-08-011-3/+4
| | |
| * | log package installationBrian C. Lane2012-08-011-3/+4
| | | | | | | | | | | | | | | log to syslog as well as /mnt/sysimage/root/install.log so that virtio and remote logging see the progress of the install.
| * | use the virtio rsyslog template (#820106)Brian C. Lane2012-08-011-1/+1
| | | | | | | | | | | | | | | The rsyslog template for forwarding logs over virtio was lost when loader was removed.
| * | keep yum cache when installing to targetBrian C. Lane2012-07-301-6/+7
| | | | | | | | | | | | | | | | | | Keep the yum cache instead of deleting it and re-downloading it. This saves time and bandwidth, and also helps reduce memory since / is on a ram backed device-mapper device.
| * | repo.mirrorlist may be NoneBrian C. Lane2012-07-271-1/+3
| | |
| * | Make sure we tear down mounted device regardless of its contents.David Lehman2012-07-271-1/+1
| | |
| * | Require new version of python-mehVratislav Podzimek2012-07-271-1/+1
| | | | | | | | | | | | We need to require a newer (Gtk3) version of python-meh.
| * | Display the hint how to quit the debuggerVratislav Podzimek2012-07-271-0/+2
| | | | | | | | | | | | | | | People usually don't know that 'continue' command quits the debugger, so display them a little hint.
| * | Just switch back to tty6 when 'c' is used in the post-mortem pdb sessionVratislav Podzimek2012-07-271-1/+4
| | | | | | | | | | | | | | | Killing ananconda left no way to report a bug via python-meh/libreport once user entered the post-mortem pdb session.
generated by cgit (git 2.34.1) at 2025-09-29 22:23:47 +0000