summaryrefslogtreecommitdiffstats
path: root/modules.d/slx-dmsetup
Commit message (Collapse)AuthorAgeFilesLines
* [slx-dmsetup] Try to get swap-service to actually run after we detected all HDDsSimon Rettberg2025-04-151-2/+4
|
* [slx-dmsetup] Tweak service ordering, wait for HDDsSimon Rettberg2025-04-092-3/+24
|
* Fix dumb typoSimon Rettberg2025-04-091-1/+1
|
* [slx-dmsetup] Remove unreachable exit statementsSimon Rettberg2025-04-091-2/+0Star
|
* [slx-dmsetup] Split dmsetup-slx-device into smaller servicesSimon Rettberg2025-04-096-198/+188Star
|
* Avoid loading dracut-lib.sh in several scriptsSimon Rettberg2025-04-021-9/+11
|
* [slx-dmsetup] More robust error handling, more loggingSimon Rettberg2025-03-111-112/+179
|
* [slx-dmsetup] Better not encrypt the rootfs AFTER building the CoW layerSimon Rettberg2024-06-141-16/+14Star
|
* [slx-dmsetup] Try to fix some snapshot brain damageSimon Rettberg2024-06-141-10/+8Star
| | | | | | | create_snapshot was always using hard-coded $writable_device, but surrounding code at one callsite suggested it was expected to use $scratch_device - this would at least explain why we could end up with overlapping dm targets... -.-
* [slx-dmsetup] busybox mktemp needs AT LEAST 6 X -.-Simon Rettberg2024-06-141-4/+6
|
* [slx-dmsetup] Fix another busybox mktemp callSimon Rettberg2024-06-141-1/+1
|
* [slx-dmsetup] Fix braindead mktemp call by removing itSimon Rettberg2024-06-131-4/+5
|
* [slx-dmsetup] Kill stray lineSimon Rettberg2024-06-131-1/+0Star
|
* [slx-dmsetup] Avoid unnecessary dm targets, rename crypted onesSimon Rettberg2024-06-131-34/+72
| | | | | | | | | | | | | | | | | | | | | | | | | 1) Avoid creating linear targets that would be fully covered by an encrypted target later. Also don't create a linear slice for the thinpool if it's going to be the only thing we put in our writable partition. 2) Previously, when we created an encrypted device, it would look like | ` devicename ` devicename-crypt Which was a bit confusing/surprising, because the final name was not what was specified in our custom partition table. Now, the result would look like | ` devicename-k ` devicename So the final device you'd use has the expected name.
* [slx-dmsetup] Make sure id44 slices are 4k-alignedSimon Rettberg2024-03-011-4/+8
|
* [slx-dmsetup] Only resize rootfs if no such service exists in stage 4Simon Rettberg2024-01-291-1/+1
|
* [slx-dmsetup] Only grow rootfs if we don't do it in stage4Simon Rettberg2024-01-292-2/+4
|
* [slx-dmsetup] grow: Preserve exit status now that it's a serviceSimon Rettberg2024-01-291-3/+2Star
|
* Try to use more systemd servicesSimon Rettberg2024-01-294-6/+32
|
* [slx-dmsetup] Use new is_on function from slx-toolsSimon Rettberg2023-07-061-1/+1
|
* [slx-dmsetup] Introduce SLX_ID44_CRYPT for encryptingSimon Rettberg2023-07-061-6/+22
| | | | | This will encrypt the entire writable device(s) before creating any partitions and pools inside it.
* [slx-dmsetup] Fix dm-crypt usage by making sure we get an encryption keySimon Rettberg2023-05-191-2/+8
|
* [slx-dmsetup] Cannot make a local var with same name as global readonlySimon Rettberg2022-08-111-3/+3
|
* [slx-dmsetup] Handle absolute slice sizes first and remember used spaceSimon Rettberg2022-08-101-2/+19
|
* [slx-dmsetup] Add linear swap space if slx.swap is in command lineSimon Rettberg2022-08-091-4/+48
| | | | .... and if we don't seem to have an existing swap partition
* [*] Redirect 'set -x' output exclusively to /run/openslx/*Simon Rettberg2022-06-101-1/+1
| | | | | | | Having normal info/warn/echo/<tool> output in journal should be preferred, so we can correlate certain events and actions with other system services and scripts, instead of having everything intermixed with the xtrace spam in a separate log file without timestamps.
* [slx-dmsetup] Properly disable future ntfsfree stuff if already activeSimon Rettberg2022-05-241-1/+4
|
* [slx-dmsetup] linear ID44 target; support SLX_NTFSFREE = "always"Simon Rettberg2022-05-231-130/+193
|
* [slx-dmsetup] reintroduce linear 1:1 mapping of scratch deviceJonathan Bauer2021-02-171-1/+13
| | | | | | | | Using the physical partition of NVMe directly led to errors when creating the pool. As a workaround, we can "wrap" it in a linear target of the same size. While this is superfluous in general, this works around that specific problem. We can remove this workaround when we have a better idea on how to fix pool creation on NVMes.
* [slx-dmsetup] remove deprecated scriptJonathan Bauer2020-11-132-34/+0Star
|
* [slx-dmsetup] copy ntfsfree within PATHJonathan Bauer2020-10-231-1/+2
|
* [slx-dmsetup] include ntfsfreeJonathan Bauer2020-10-221-1/+3
| | | | and remove old comment that was wrong anyway
* [slx-dmsetup] Don't fall back to /run/openslx for CoW-tmpfsSimon Rettberg2020-10-091-5/+2Star
| | | | This would mask anything already in /run/openslx.
* [slx-dmsetup] Minor cleanupSimon Rettberg2020-10-091-4/+3Star
|
* [slx-dmsetup] part-table: Allow different units for min and maxSimon Rettberg2020-10-021-32/+51
| | | | | It can useful to specify a minumum size in absolute terms, but have the maximum value be a percentage of the writable device.
* [slx-dmsetup] Add overprovisioning formula for CoW volumeSimon Rettberg2020-10-021-12/+16
|
* [slx-dmsetup] Use _sz for sizes expressed in 512b sectorsSimon Rettberg2020-10-011-59/+61
|
* [slx-dmsetup] Avoid creating linear targets that are 1:1 mappingsSimon Rettberg2020-10-011-87/+120
| | | | Also a bit more cleanup and comments.
* [slx-dmsetup] Add support for ntfsfreeSimon Rettberg2020-09-301-2/+56
| | | | | | | | | | | | For now, prepare for later on-demand growing of the pool-data davice by using unused space of NTFS partitions. We do this by scanning available NTFS partitions and recording the available free space on them. Additionally, we make the pool-metadata volume a little larger to account for potential future growth of the data part. This feature is controlled by the SLX_NTFSFREE variable which can be empty/"never", "backup" for the scenario described above, or "always" for always using NTFS partition's free space (to be implemented).
* [slx-dmsetup] Block zeroing not required in our use caseSimon Rettberg2020-09-301-1/+1
| | | | | | This is more useful if you pass volumes from the thin pool to potentially untrused VMs or containers. We just run a normal file system on top which takes care of zeroing for us already.
* [slx-dmsetup] Fix another dd callSimon Rettberg2020-09-301-1/+1
|
* [slx-dmsetup] Simplifications, commentsSimon Rettberg2020-09-301-21/+33
|
* [slx-dmsetup] Fix minor coding style issues, fix dd callSimon Rettberg2020-09-291-28/+37
| | | | | Change low watermark from 4 to 100MB. Add a few comments. dd call was somehow working, but was reading from a random stdin?
* [slx-dmsetup] simplier random dm-crypt key generationJonathan Bauer2020-09-171-1/+1
|
* [slx-dmsetup] explicitely add aes-xts-plain64 related kernel modulesJonathan Bauer2020-09-151-5/+4Star
|
* [slx-dmsetup] remove unneeded fallocateJonathan Bauer2020-05-251-1/+1
|
* [slx-dmsetup] more work towards persistent thin poolJonathan Bauer2020-05-251-15/+11Star
|
* Merge branch 'nobash-merge' into downloader-nobash-mergeJonathan Bauer2020-05-255-180/+62Star
|
* [slx-dmsetup] fallback to RAMdisk pool metadataJonathan Bauer2020-05-141-9/+11
|
* [slx-dmsetup] fix noop in if/thenJonathan Bauer2020-05-141-0/+2
|
generated by cgit v1.2.3-55-g7522 (git 2.39.0) at 2026-03-02 17:51:43 +0100