Am 23.03.13 18:17, schrieb poma: > On 23.03.2013 06:05, Harald Hoyer wrote: >> Am 22.03.2013 21:32, schrieb poma: >>> On 22.03.2013 13:00, Harald Hoyer wrote: >>> ... >>>> you must not resume from a raid array >>>> >>> >>> Let me show you how country feels :) >>> >>> Version: 025-35.git20130211.fc18 [1] >>> >>> /etc/dracut.conf.d/dracut-custom.conf: >>> # Custom dracut config file >>> >>> # /etc/dracut.conf.d/*.conf will override the settings in /etc/dracut.conf >>> >>> # build initrd only to boot current hardware >>> hostonly="yes" >>> >>> # Bring up <device> in initramfs, <device> should be the device name. >>> # Beside device name, it would be nice if one can use fstab alike UUID. >>> # This can be useful in hostonly mode for resume support when >>> # your swap is on LVM an encrypted partition or on a MD RAID partition. >>> add_device+=" /dev/md1 " >>> .. >>> >>> The corresponding command: >>> dracut --hostonly --device /dev/md1 [-f] [-v] >>> .. >>> >>> PM: Hibernation image not present or could not be loaded. >>> PM: Hibernation mode set to 'platform' >>> PM: Marking nosave pages: [mem 0x0009e000-0x000fffff] >>> PM: Marking nosave pages: [mem 0xcfef0000-0xffffffff] >>> PM: Marking nosave pages: [mem 0xc4000000-0xc7ffffff] >>> PM: Basic memory bitmaps created >>> PM: Syncing filesystems ... done. >>> PM: Preallocating image memory... done (allocated 279164 pages) >>> PM: Allocated 1116656 kbytes in 0.30 seconds (3722.18 MB/s) >>> PM: freeze of devices complete after 1463.281 msecs >>> PM: late freeze of devices complete after 0.472 msecs >>> PM: noirq freeze of devices complete after 0.354 msecs >>> PM: Saving platform NVS memory >>> PM: Creating hibernation image: >>> PM: Need to copy 292982 pages >>> PM: Normal pages needed: 292982 + 1024, available pages: 738735 >>> PM: Restoring platform NVS memory >>> PM: noirq restore of devices complete after 21.957 msecs >>> PM: early restore of devices complete after 0.352 msecs >>> PM: restore of devices complete after 1290.944 msecs >>> PM: Image restored successfully. >>> PM: Basic memory bitmaps freed >>> .. >>> >>> Swap redundancy is good, right? ;) >>> >>> poma >>> >> >> Sure, if it works for you :) >> >> One recommendation: please use UUID= instead of relying on the kernel numbering >> scheme for devices, if you want to avoid problems. >> > > Yeah, UUIDs are being used from the day one. > But what I meant is: > add_device+=" <PARTUUID> " directive in addition to the " <device> " as > device name such as in the current 'dracut.conf' schema. > Although, this is resolved by an explicit specification of a swap UUID, > i.e. rd.md.uuid=<MD array UUID> to the boot loader, so that directive is > not required in this case. > > > poma > > /dev/disk/by-uuid/.... can be used instead of UUID= -- users mailing list users@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines Have a question? Ask away: http://ask.fedoraproject.org