v0.6.1
Highlights
- Storage-Accelerated Copy across several arrays: ArrayCreds are now chosen per VM, so one migration plan can cover VMs on different arrays, including a single VM whose disks sit on more than one array.
- Hot migrations are more stable during Periodic Sync:
- A intermittent vCenter error no longer marks a running migration Failed.
- A new stall watchdog prevents changed-block copy from staying unresponsive forever.
- vJailbreak Accelerated Copy (Hot-Add) is more reliable: Concurrent migrations never compete for a port, avoiding a race condition during concurrent migrations.
- vAssessment (beta): a separate appliance for pre-migration readiness checks, with a web UI, scoping, readiness groups and diagnostics. Beta version available as OVA.
Bug Fixes
- Fixed the Convert phase failing with “no operating system found”: each disk is now checked for readiness after it is attached to the vJailbreak VM, before virt-v2v runs.
- Fixed qemu-nbd not being confirmed as listening after an NBD server restart.
- Fixed Retry silently changing Standard Copy to vJailbreak Accelerated Copy.
- Fixed false subnet-mismatch warnings for multi-NIC VMs on different networks. Persist source network interfaces is no longer greyed out for these VMs.
- Fixed ArrayCreds rejecting valid volume types when a Cinder backend has more than one volume type.
- Fixed issues for Windows dynamic disk (LDM) migrations.
Security
- Rebuilt the Hot-Add proxy OVA to fix the sshole vulnerability (CVE-2026-123158).
Known Limitations
- Hot-Add stale snapshots: snapshots left by failed Hot-Add migrations can pile up and cause UUID conflicts. Remove them in vCenter before retrying.
- SUSE Linux Enterprise HPC 15 SP5 is not supported: virt-v2v doesn’t recognise
sle_hpc. - Users upgrading in-place from v0.5.0 or v0.6.0 to this version may hit a known issue during upgrade. Retrying the upgrade should resolve it automatically.
What’s Changed
- fix(upgrade): retry rejected deletes during cleanup so a terminating ProxyVM no longer blocks the upgrade by @sarika-pf9
- fix(sam): resolve ArrayCreds per VM and support VMs spanning several arrays by @spai-p9
- fix(controller): stop losing the Migration retryable status so RDM VMs never show Retry by @sarika-pf9
- fix(controller): stop failing running shared-RDM migrations and their plan when a sibling owner VM fails by @sarika-pf9
- fix(controller): agent reprovision stuck in VMCreated — wait for Nova delete by @spai-p9
- Fix Hot-Add proxy disk leak and false Failed migration status during periodic sync by @sarika-pf9
- Readiness checks and retry mechanism before copy stage by @meghansh-pf9
- Migration form IP warning not updated after grid refresh by @AbhijeetThakur
- Stall watchdog in periodic sync routine by @meghansh-pf9
- fix: use the git tag as release version instead of 99.99.99 by @archit-pf9
- feat: log vJailbreak version on startup in all components by @archit-pf9
- Binding v2v packages to local rpms by @meghansh-pf9
- Persist source network interfaces gating per networking spec by @sarika-pf9
- fix(v2v-helper): start Hot-Add qemu-nbd with —fork and a bind-retry loop by @spai-p9
- Retry: keep the failed migration’s storage copy method by @sarika-pf9
- Retry - The missing-VDDK block isn’t applied by @AbhijeetThakur
- Adding readiness check for each disk attached to the vJailbreak VM by @meghansh-pf9
- fix(arraycreds): accept any volume type configured for a Cinder backend by @spai-p9
- fix: fall back to a healthy Cinder host when the pinned one is unavailable by @spai-p9
- feat(claude): PR reviews check for regressions and cross-feature impact by @sarika-pf9
Docs (platform9/vjailbreak): deployment architecture best practices for nested virtualization and SAM, and a workaround for NFC memory exhaustion, both by @OmkarDeshpande7