summaryrefslogtreecommitdiff
path: root/debian/initramfs
Commit message (Collapse)AuthorAge
* removed references to /var/lib/mdadm and cleaned up scripts a bitMichael Tokarev2014-10-04
|
* remove references to old (2.x and 1.x) mdadm versionsMichael Tokarev2014-09-27
|
* create /etc/modprobe.d/mdadm.confMichael Tokarev2014-09-20
|
* denote md init errors in initramfs from fatal to warning (#733574)Michael Tokarev2014-09-20
|
* copy 64-md-raid-assembly.rules to initramfs too (#678691)Michael Tokarev2014-09-20
|
* use 63-md-raid-arrays.rules instead of old 64-md-raid.rules (#726237)Michael Tokarev2013-10-14
|
* use wait_for_udev in initramfs scriptMichael Tokarev2013-03-05
|
* add support for mdmon based on changed by Miquel van Smoorenburg (#684708)Michael Tokarev2012-10-28
|
* fix initramfs-script config name (slipped in previous changes)Michael Tokarev2012-08-25
|
* fixes for initramfs integration (#644389, #678262, #685161)Michael Tokarev2012-08-23
|
* Copy local administrator's modified udev rules as well as the system one.Dmitrijs Ledkovs2012-06-25
|
* Remove compatability with acient initramfs-tools.Dmitrijs Ledkovs2012-05-25
| | | | | | | | | On one hand, it is saver to have it in. On the other hand, people should have transitions to initramfs since etch, and it would be challenging to support acient initrd's. I'd want to see this go...
* initamfs/hook: make sure configuration file exists before accessing it.Michael Prokop2010-09-02
| | | | | | | Debian bug: #589836 Signed-off-by: Michael Prokop <mika@grml.org> Signed-off-by: martin f. krafft <madduck@debian.org>
* Match UUID case-insensitivelyMario 'BitKoenig' Holbe2010-05-28
| | | | | | | while checking for running arrays not listed in mdadm.conf Debian-bug: #583545 Signed-off-by: martin f. krafft <madduck@debian.org>
* translate between /dev/mdX and /dev/md/Xmartin f. krafft2010-01-29
| | | | Signed-off-by: martin f. krafft <madduck@debian.org>
* check for running devices based on UUIDmartin f. krafft2010-01-29
| | | | Signed-off-by: martin f. krafft <madduck@debian.org>
* Do not single-quote homehost in initramfs scriptmartin f. krafft2010-01-27
| | | | | Debbug: 549083 Signed-off-by: martin f. krafft <madduck@debian.org>
* Copy udev rules to initramfsmartin f. krafft2009-10-27
| | | | | | | | | | | udev stopped doing that, so we have to. Debian-Bug: http://bugs.debian.org/549298 Signed-off-by: martin f. krafft <madduck@debian.org> Conflicts: debian/changelog
* only store MD_HOMEHOST if config does notmartin f. krafft2009-07-21
| | | | Signed-off-by: martin f. krafft <madduck@debian.org>
* Remove auto-update-homehost flagmartin f. krafft2009-07-21
| | | | | | | | | The flag was removed in 3.0. It also should not be needed anymore, since we don't really want to update the homehost field anyway, and the 3.0 heuristics, which assembles non-homehost arrays as read-only, is fine. Debian bug: #537820 Signed-off-by: martin f. krafft <madduck@debian.org>
* Make initramfs script exit after printing the prerequisitesmartin f. krafft2009-05-03
| | | | Signed-off-by: martin f. krafft <madduck@debian.org>
* Remove extra -x check for mdadm binaryCristian Ionescu-Idbohrn2009-04-27
| | | | Signed-off-by: martin f. krafft <madduck@debian.org>
* Remove extra characters from variable initialisationsCristian Ionescu-Idbohrn2009-04-27
| | | | Signed-off-by: martin f. krafft <madduck@debian.org>
* Properly quote argument to prereq queryCristian Ionescu-Idbohrn2009-04-26
| | | | Signed-off-by: martin f. krafft <madduck@debian.org>
* Fix mdadm.conf check for know arraysmartin f. krafft2009-04-26
| | | | | | | | | | | | Fix the check whether mdadm.conf defines all devices known to the system: we were checking against $devs earlier, which is only those devices required to boot; $devpairs has them all. Thanks Cristian Ionescu-Idbohrn. Debian bug: 525655 Signed-off-by: martin f. krafft <madduck@debian.org>
* Check if all known arrays are listed in mdadm.confmartin f. krafft2009-04-25
| | | | | | | | | Patch from Adrian Bridgett: the initramfs hook now checks to see if all known arrays are listed in mdadm.conf and issues a warning if this is not the case. Debian bug: 519328 Signed-off-by: martin f. krafft <madduck@debian.org>
* remove trailing whitespacemartin f. krafft2009-03-09
|
* Make initramfs script depend on multipath to ensure its script is run before ↵martin f. krafft2009-02-23
| | | | ours (closes: #516605).
* record snapshotting upstream in changelogmartin f. krafft2009-02-16
| | | | Signed-off-by: martin f. krafft <madduck@debian.org>
* use udevsettle to wait for devicesmartin f. krafft2008-07-02
|
* remove prereq on udev_helpermartin f. krafft2008-07-02
|
* update copyright year to 2008martin f. krafft2008-07-02
|
* make initramfs script wait for root linkmartin f. krafft2008-07-02
| | | | | if $ROOT is /dev/disk/by-id/md-*, then we wait up to 10 seconds for it to appear because udev may be slow...
* store md.conf file in conf.d subdir in initramfsmartin f. krafft2008-07-02
|
* remove mdrun completelymartin f. krafft2008-07-02
|
* handle module name synonyms in initramfs script (#432585)martin f. krafft2007-10-09
| | | | | | | | mdadm's initramfs hook script previously expected module names to be of the form raidX and ignored synonyms, like X, which other parts of mdadm treat just fine. This patch maps the synonyms to the module names to use. Closes Debian bug #432585
* cleanupmartin f. krafft2007-10-09
|
* * Use short option in initramfs script mkdir call, so make it klibc-utilsmartin f. krafft2007-10-09
| | | | compatible; thanks maximilian attems for the patch (closes: #443436).
* add --scan to the single-device-assembly-codepath in the initramfs local-top ↵martin f. krafft2007-10-09
| | | | | | script debbug 440703
* De-escalate the initramfs hook warning about an "emergency procedure"martin f. krafft2007-10-09
| | | | | | | | Instead, simply note that this involves initramfs assembling arrays it finds automatically at boot. Also added a FAQ entry on how to turn off the init script warning when no arrays are found debbug 434934
* * Does some sanity checking for proper format of level= arguments inmadduck2007-10-07
| | | | | | | mdadm.conf and bails if an error is found. Since the RAID levels are used verbatim as module names, a discrepancy might lead to an unusable system. Instead, we thus use the emergency fallback. Thanks to Andrew Sackville-West for spotting this and helping me with the fix.
* bump copyright year to 2007madduck2007-10-07
|
* * Release to unstable.madduck2007-05-05
| | | | | | | * Start arrays read-only in initramfs to prevent syncing and hence enable resuming/freezing. The arrays will automatically sync as soon something writes to it; thanks to Tim Dijkstra, Neil Brown, and Luis Rodrigo Gallardo Cruz (closes: #415441).
* * Execute udevtrigger after assembly of arrays during initramfs processing;madduck2007-03-31
| | | | thanks to Michael Prokop (closes: #416658).
* * Hard-code path to /sbin/mdadm binary rather than searching the $PATH. Thismadduck2006-12-19
| | | | | closes: #403307 and should be a little more robust in the presence of installations of mdadm in /usr/local.
* * Only parse ARRAY lines from configuration file when collecting the arraymadduck2006-12-08
| | | | | | | pairs. Thanks to Daniel Dehennin for the bug report and suggested fix (closes: #402106). * Prevent modules from being loaded during initramfs time if no arrays are to be assembled at this stage.
* * Prevent initramfs hook from exiting prematurely when VERBOSE=false.madduck2006-11-13
|
* don't forget to create the target directorymadduck2006-11-11
|
* * Actually remove mdadm.conf on purge; previously, the generation ofmadduck2006-11-11
| | | | | a temporary file for initramfs would screw up the purging; thanks to Fabrice Lorrain for the report (closes: #398088).
* write to stderr instead of stdoutmadduck2006-10-29
|