-
Notifications
You must be signed in to change notification settings - Fork 400
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
LVM boot failure with systemd-253 #2211
Comments
It appears this issue will be fixed at the systemd level as a patch was produced over there. Sorry for the noise. |
Reopening as the systemd devs seem to think dracut needs fixing anyway. |
There's a good explanation here: systemd/systemd#26488 (comment)
EDIT: the file cannot be simply created in the first place, due to sandboxing. |
I think we should move generator_wait_for_dev to a cmdline hook. It is not the nicest solution, but logically it belongs there. |
@lnykryn it's more of an workaround not a solution and afaikt would only solve half the problem. |
Why is not a solution to only move the part of dracut/modules.d/90crypt/parse-crypt.sh Lines 169 to 174 in aca5120
The only missing part would be how to know from the cmdline hook the directory that systemd passes to generators (i.e. usually /run/systemd/generator), but we can hardcode it for now. |
…or dir Altough it was already documented in systemd.generator(7) that generators must not write to locations other than those passed as arguments, since systemd/systemd@ca6ce62d systemd executes generators in a mount namespace "sandbox", so now the hooks created by the rootfs-generator are lost. These hooks are created using the root= cmdline argument, so this patch moves the creation of these hooks to a cmdline hook. Fixes issue dracutdevs#2211 Fixes issue dracutdevs#2225
This issue will affect openSUSE Tumbleweed soon, systemd will be updated from v252.6 to v253.1 and it will not carry the Fedora workaround, so I'd like to have the dracut upstream fix agreed. @lnykryn I tried your proposal and it seems to work fine (#2225 is fixed). Do you plan to submit a PR? If not, I can send it. |
@aafeijoo-suse Please file the PR. |
…or dir Although it was already documented in systemd.generator(7) that generators must not write to locations other than those passed as arguments, since systemd/systemd@ca6ce62d systemd executes generators in a mount namespace "sandbox", so now the hooks created by the rootfs-generator are lost. These hooks are created using the root= cmdline argument, so this patch moves the creation of these hooks to a cmdline hook. Fixes issue dracutdevs#2211 Fixes issue dracutdevs#2225
…or dir Although it was already documented in systemd.generator(7) that generators must not write to locations other than those passed as arguments, since systemd/systemd@ca6ce62d systemd executes generators in a mount namespace "sandbox", so now the hooks created by the rootfs-generator are lost. These hooks are created using the root= cmdline argument, so this patch moves the creation of these hooks to a cmdline hook. Fixes issue dracutdevs#2211 Fixes issue dracutdevs#2225
…or dir Although it was already documented in systemd.generator(7) that generators must not write to locations other than those passed as arguments, since systemd/systemd@ca6ce62d systemd executes generators in a mount namespace "sandbox", so now the hooks created by the rootfs-generator are lost. These hooks are created using the root= cmdline argument, so this patch moves the creation of these hooks to a cmdline hook. Fixes issue dracutdevs#2211 Fixes issue dracutdevs#2225
…or dir Although it was already documented in systemd.generator(7) that generators must not write to locations other than those passed as arguments, since systemd/systemd@ca6ce62d systemd executes generators in a mount namespace "sandbox", so now the hooks created by the rootfs-generator are lost. These hooks are created using the root= cmdline argument, so this patch moves the creation of these hooks to a cmdline hook. Fixes issue dracutdevs#2211 Fixes issue dracutdevs#2225
https://build.opensuse.org/request/show/1070393 by user fbui + dimstar_suse - Add 5000-core-manager-run-generators-directly-when-we-are-in-.patch, a temporary workaround until dracutdevs/dracut#2211 is fixed in dracut. - Upgrade to v253.1 (commit 6c327d74aa0d350482e82a247d7018559699798d) See https://github.com/openSUSE/systemd/blob/SUSE/v253/NEWS for details. * Rebased 0001-conf-parser-introduce-early-drop-ins.patch * Ship systemd-journald-audit.socket again: it can now be disabled via the usual "systemctl disable" mechanism to stop collection of audit messages. Note that it's handled by the preset logic, which turns it off by default. * TEST_06_SELINUX needs selinux-policy-devel.
https://build.opensuse.org/request/show/1072270 by user fbui + dimstar_suse - Add 5002-systemctl-explicitly-cast-the-constants-to-uint64_t.patch (bsc#1209305) Added temporarily until it's merged in either the stable v253 branch or in the SUSE git repo. - Add 5001-Revert-core-propagate-stop-too-if-restart-is-issued.patch until systemd/systemd#26839 is fixed properly. - testsuite: mtools is required by TEST-58-REPART - testsuite: swtpm and tpm2.0-tools are needed by TEST-70-TPM2 - Add 5000-core-manager-run-generators-directly-when-we-are-in-.patch, a temporary workaround until dracutdevs/dracut#2211 is fixed in dracut. - Upgrade to v253.1 (commit 6c327d74aa0d350482e82a247d7018559699798d) See https://github.com/openSUSE/systemd/blob/SUSE/v253/NEWS for details. * Rebased 0001-conf-parser-introduce-e
lvm with latest systemd and dracut doesn't boot up. Upstream issues for reference: - dracut: dracutdevs/dracut#2211 - systemd: systemd/systemd#26488 Temporary patch applied until this issue is resolved systemd/systemd#26494 See systemd/systemd#26488 (comment) Change-Id: I1964ff2a82b7c2d6c53e02b4368a8d5d6f9a03a1 Signed-off-by: Tapas Kundu <[email protected]> Reviewed-on: http://photon-jenkins.eng.vmware.com:8082/20134
lvm with latest systemd and dracut doesn't boot up. Upstream issues for reference: - dracut: dracutdevs/dracut#2211 - systemd: systemd/systemd#26488 Temporary patch applied until this issue is resolved systemd/systemd#26494 See systemd/systemd#26488 (comment) Change-Id: I1964ff2a82b7c2d6c53e02b4368a8d5d6f9a03a1 Reviewed-on: http://photon-jenkins.eng.vmware.com:8082/20132 Tested-by: gerrit-photon <[email protected]> Reviewed-by: Susant Sahani <[email protected]>
…or dir Although it was already documented in systemd.generator(7) that generators must not write to locations other than those passed as arguments, since systemd/systemd@ca6ce62d systemd executes generators in a mount namespace "sandbox", so now the hooks created by the rootfs-generator are lost. These hooks are created using the root= cmdline argument, so this patch moves the creation of these hooks to a cmdline hook. Fixes issue #2211 Fixes issue #2225
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
This fixes LVM2 volume group initialisation with systemd 254.4. Ref: dracutdevs/dracut#2211
https://build.opensuse.org/request/show/1070393 by user fbui + dimstar_suse - Add 5000-core-manager-run-generators-directly-when-we-are-in-.patch, a temporary workaround until dracutdevs/dracut#2211 is fixed in dracut. - Upgrade to v253.1 (commit 6c327d74aa0d350482e82a247d7018559699798d) See https://github.com/openSUSE/systemd/blob/SUSE/v253/NEWS for details. * Rebased 0001-conf-parser-introduce-early-drop-ins.patch * Ship systemd-journald-audit.socket again: it can now be disabled via the usual "systemctl disable" mechanism to stop collection of audit messages. Note that it's handled by the preset logic, which turns it off by default. * TEST_06_SELINUX needs selinux-policy-devel.
https://build.opensuse.org/request/show/1072270 by user fbui + dimstar_suse - Add 5002-systemctl-explicitly-cast-the-constants-to-uint64_t.patch (bsc#1209305) Added temporarily until it's merged in either the stable v253 branch or in the SUSE git repo. - Add 5001-Revert-core-propagate-stop-too-if-restart-is-issued.patch until systemd/systemd#26839 is fixed properly. - testsuite: mtools is required by TEST-58-REPART - testsuite: swtpm and tpm2.0-tools are needed by TEST-70-TPM2 - Add 5000-core-manager-run-generators-directly-when-we-are-in-.patch, a temporary workaround until dracutdevs/dracut#2211 is fixed in dracut. - Upgrade to v253.1 (commit 6c327d74aa0d350482e82a247d7018559699798d) See https://github.com/openSUSE/systemd/blob/SUSE/v253/NEWS for details. * Rebased 0001-conf-parser-introduce-e
Describe the bug
Updating to systemd-253 causes boot failure when dracut is used to generate the initrd and the root volume is LVM.
Distribution used
Arch Linux
Dracut version
Arch is on dracut 056-1 but another user has tested the current dracut git with same result.
Init system
systemd
To Reproduce
Expected behavior
Boot success.
Additional context
The offending systemd commit has been bisected. Please see:
systemd/systemd#26488
Additional context in this Arch bug report:
https://bugs.archlinux.org/task/77569
At this stage it seems like dracut is not coping with latest systemd changes. Thanks for any help.
The text was updated successfully, but these errors were encountered: