Skip to content

Adding testcase for kdump after migration! - #36

Open
Anushree-Mathur wants to merge 1 commit into
lop-devops:KVMCIfrom
Anushree-Mathur:migration_kdump
Open

Adding testcase for kdump after migration!#36
Anushree-Mathur wants to merge 1 commit into
lop-devops:KVMCIfrom
Anushree-Mathur:migration_kdump

Conversation

@Anushree-Mathur

Copy link
Copy Markdown

Test scenario:

  1. Configure VM with host-model CPU (auto-detected on ppc64/ppc64le) and set on_crash=restart so the guest reboots after a kernel panic.
  2. Perform live migration to the destination host.
  3. Open a serial console session on the destination guest — serial is used instead of SSH because the VM MAC is no longer reachable on the source host virbr0 after migration, which causes wait_for_login() to fail due to stale ARP/address-cache entries.
  4. Record pre-crash vmcore list from /var/crash/.
  5. Trigger a kernel panic via echo c > /proc/sysrq-trigger.
  6. Wait for the guest to reboot (on_crash=restart).
  7. Re-open serial session, collect post-crash vmcore list and confirm at least one new vmcore was created — failure is reported otherwise.

Config additions:

  • variants kdump: kdump_migration (post_migration_kdump=yes) / no_kdump
  • Additional migration modes: p2p_tunneled, postcopy, postcopy_p2p, auto_converge, auto_converge_p2p — kdump crosses all of them.
  • custom cpu_mode excludes ppc64le,ppc64 (no named model available).

Signed-off-by: Anushree-Mathur <anushree.mathur@linux.ibm.com>

Test scenario:
1. Configure VM with host-model CPU (auto-detected on ppc64/ppc64le)
   and set on_crash=restart so the guest reboots after a kernel panic.
2. Perform live migration to the destination host.
3. Open a serial console session on the destination guest — serial is
   used instead of SSH because the VM MAC is no longer reachable on
   the source host virbr0 after migration, which causes wait_for_login()
   to fail due to stale ARP/address-cache entries.
4. Record pre-crash vmcore list from /var/crash/.
5. Trigger a kernel panic via echo c > /proc/sysrq-trigger.
6. Wait for the guest to reboot (on_crash=restart).
7. Re-open serial session, collect post-crash vmcore list and confirm
   at least one new vmcore was created — failure is reported otherwise.

Config additions:
- variants kdump: kdump_migration (post_migration_kdump=yes) / no_kdump
- Additional migration modes: p2p_tunneled, postcopy, postcopy_p2p,
  auto_converge, auto_converge_p2p — kdump crosses all of them.
- numa_cell updated to 32 vCPUs / 16 GiB to match a Power LPAR layout.
- custom cpu_mode excludes ppc64le,ppc64 (no named model available).
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant