EL 10 Migration via Ascender
Applies to: RHEL 10.0, migrating to RLC Pro 10. The
rlc-protier is the only tier available on EL10.
Overview
This guide covers deploying the migrate2rlc Ansible playbook through CIQ's Ascender automation platform for a RHEL 10 fleet. The Ascender mechanics do not change for EL10. Project, credential, inventory and Job Template setup are the same steps the EL9 guide walks through, so read EL 9 via Ascender for those and for the Standalone Ansible alternative. What this guide carries is everything that differs on EL10, which is mostly what the play refuses.
The Job Template
Same shape as the EL9 template, with one thing to hold to. In Extra Variables:
accept_disclaimer: true
depot_tier: rlc-pro
depot_tier: rlc-pro is not a default you can quietly override. The play
carries an assert named Validate EL10 supports the rlc-pro tier only, and
a job that reaches an EL10 host with any other tier fails that host before
any migration step, before any package swap, repository rewrite or
branding, with the message EL10 currently supports the rlc-pro tier only. rlc-plus-10 does exist as a depot channel, so the
refusal is not about a missing channel. Nothing brands an EL10 host for the
Plus tier and no EL10 Plus migration has been tested end to end, and a
part-branded host is the one state this tool refuses to produce. The same
assert fires on a host already running RLC Pro 10, so there are no tier
changes on EL10 either.
Failing the assert does not mean the host is untouched. A job that also
requested a backup (backup_mode) has taken it before the assert runs, so
a failed host carries the backup archive and any helper packages the
backup installed, the NFS client, cifs-utils or rsync. The backup
block unmounts its share on the way out. Nothing of the migration itself
has run.
migration_tier is accepted as an alias for depot_tier. Set one or the
other, not both.
Everything a Job Template supplies through Extra Variables or a credential injector arrives as an Ansible extra var, and extra vars cannot be overridden from inside the playbook. That is what makes the Depot credential injector work, and it is also why a wrong tier in a survey cannot be corrected anywhere later in the run. Fix the survey and relaunch.
EL10-Specific Notes
- Managed-node Python needs no action on a stock RHEL 10 image. Its
/usr/bin/python3is new enough for every ansible-core floor the playbook derives. The first play checks anyway and stops with instructions if a minimised or hardened image has had the interpreter removed. Unlike EL8, this is not something to pre-stage across a fleet. If pre-staging Python is ever the blocker, the bash scriptmigrate2rlc.shhas no interpreter prerequisite at all. - FIPS mode: the play reads the kernel's own flag at
/proc/sys/crypto/fips_enabledand fails the host if it reads1. EL10 does not shipfips-mode-setup, the tool EL9 uses, so there is no command the playbook can suggest. Disable FIPS by the procedure your RHEL 10 documentation gives before you launch the job. - CentOS Stream 10 is refused with its own task, before any migration step, because Stream 10 repositories moved to mirror.stream.centos.org and the tool has not been updated or tested for that layout. A job that requested a backup has taken it by then too, leaving what the tier-assert note above describes.
- Air-gapped migration is not available on EL10.
airgapped=trueon an EL10 host fails with the air-gap major list, which is 8 and 9 only. - GPG keys: the Rocky 10 key and the CIQ content key for EL10 are embedded
in the playbook and verified against pinned SHA512 digests. The CIQ
content key is a different key on EL10, CIQ Signing Key RSA (RLC 10),
than the V2 key EL8 and EL9 content uses. If CIQ rotates the content key,
the run stops and prints the observed digest; confirm it against CIQ and
pass it back with
-e accept_ciq_key_sha512=DIGESTat launch time. It accepts that one digest and nothing else. - Kernel: the playbook installs the CIQ kernel explicitly on aarch64 EL9
only, so on EL10 the host finishes on whichever
kernel-coreis newest, which on x86_64 is currently the stock Rocky kernel. That is the supported outcome, and the end-of-run kernel report names what booted. - Secure Boot: the boot-chain replacement measured on EL9 x86_64 has not been measured on EL10, and no claim is made about it. Snapshot the fleet, or stage the job in batches with console access available, before running it against production EL10 UEFI hosts.
- Tier release package:
rlc-pro-releaseexists for EL10 and is installed by the play, the same package EL9 Pro uses.verify_migration.shasserts the landed tier from it.
Verification runs, and a failure fails the job
Everything the EL9 guide describes here is unchanged on EL10. The play
copies verify_migration.sh to each target and runs it from the post-reboot
recovery service, and the final task fails the host when that script
reported any FAIL: line. The verdict fires only when auto_reboot is
true, because verification runs after the reboot. With the default
auto_reboot: false the job ends with a NOT YET VERIFIED notice instead,
and you grade the fleet yourself by reading
/var/lib/migrate2rlc/migration-summary.txt on each host. For fleet work,
setting auto_reboot: true in the Job Template is what makes the job's own
red/green meaningful.
The post-reboot recovery step also removes the pre-migration kernel
packages, guarded so the running kernel and the last installed kernel-core
are never removed, re-sweeps for boot entries orphaned by that removal, and
regenerates the rescue kernel and initramfs. This is destructive, and it
removes a fallback a fleet operator may be counting on. Once the old
kernel packages are gone, recovery is through the rescue entry, console
access or a snapshot, not the GRUB menu. See Post-Migration Kernel and Boot
Cleanup in the root README for the full detail.
Telemetry and the Opt-In Survey
Unchanged from the other EL guides. After migration, the post-reboot
recovery step reports two facts automatically, with no consent prompt,
whether the migration succeeded (migration_outcome) and whether the
target ended up enrolled in an RLC Pro repository
(rlc_pro_repo_configured). Nothing else is sent. The destination defaults
to empty, so no network call happens today. Add -e migrate2rlc_telemetry_opt_out=true to opt out, or -e migrate2rlc_survey_consent=true (with -e migrate2rlc_survey_failure_note
for the note) to opt in to the local-only survey. See Telemetry and the
Opt-In Survey in the root README.
Monitoring and Troubleshooting
The EL9 guide's monitoring notes apply unchanged. The EL10 entries:
- The assert
Validate EL10 supports the rlc-pro tier onlyfails a host: the job reached an EL10 host withdepot_tierset torlc-plusorrlc-96-lts. Nothing of the migration ran on that host, though a job that also setbackup_modehad already taken its backup and installed any helper packages it needed. Relaunch withdepot_tier: rlc-pro, or keep those hosts on EL9. CentOS Stream 10 is not a supported migration source: no workaround in this release.- Air-gap refusal naming majors 8 and 9: EL10 cannot migrate air-gapped. Migrate online.
- FIPS refusal on start: the play read
1from/proc/sys/crypto/fips_enabled. Disable FIPS on that host, reboot, and relaunch the job against it. - The play fails the last task on a
FAIL: Tier mismatchline: the host landed a different tier than the job requested. This is a report, not a rescue. On EL10 the recovery is a re-run torlc-proor a restore from the pre-migration backup.
Standalone Ansible Alternative
Same as EL9. Put the token in a file on the control node and point
depot_token_file at it, and keep depot_tier at rlc-pro:
install -m 600 /dev/null ~/.ciq-token
vi ~/.ciq-token
ansible-playbook migrate2rlc.yml \
-i target_host, \
-e accept_disclaimer=true \
-e depot_tier=rlc-pro \
-e depot_username=YOUR_USER \
-e "depot_token_file=$HOME/.ciq-token"
Full rules for the token, the two-files requirement and the extra variables are in Standalone Ansible and the RHEL 10 bash guide.
DISCLAIMER
The migrate2rlc script and playbook are provided "as is" without warranty of any kind, express or implied. Use of these tools is entirely at your own risk. CIQ, Inc. assumes no responsibility or liability for any data loss, system downtime, service disruption, or other damage resulting from the use or misuse of these tools. It is the user's responsibility to ensure adequate backups and testing in a non-production environment before performing any migration on production systems.
This tool automatically reports migration_outcome (success/failure) and
whether the RLC Pro repo ended up configured, and nothing else. Add -e migrate2rlc_telemetry_opt_out=true to opt out. Add -e migrate2rlc_survey_consent=true to additionally share your source
distribution and an optional failure note, captured to a local file only
and never sent unless you attach it to a support ticket yourself.