From 0484b5d7c6c11f6ae832bf504d528e04d5fa31aa Mon Sep 17 00:00:00 2001
From: Denys Fedoryshchenko <denys.f@collabora.com>
Date: Wed, 10 Jun 2026 15:57:46 +0300
Subject: ci: harden s390x qemu test against modloop/disk flakes

The s390x netboot initramfs ships virtio_net but not virtio_blk; the
block driver only becomes available after the modloop image is
downloaded over the network at boot. When that download fails, sshd
still comes up but /dev/vda never appears and setup-disk dies with
"/dev/vda is not a block device suitable for partitioning".

- Wait for /sys/block/vda/device before running setup-alpine,
  restarting the modloop and hwdrivers services on each retry.
- Log the QEMU console (screen -L) for both VM launches and dump the
  logs at the end of the job, so failures of the second boot (the
  other recurring flake, where ssh never comes back after reboot)
  are debuggable.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---
 .github/workflows/run-tests-bigendian.yml | 25 +++++++++++++++++++++++--
 1 file changed, 23 insertions(+), 2 deletions(-)

(limited to '.github')

diff --git a/.github/workflows/run-tests-bigendian.yml b/.github/workflows/run-tests-bigendian.yml
index eb09aa9f..389cd2a7 100644
--- a/.github/workflows/run-tests-bigendian.yml
+++ b/.github/workflows/run-tests-bigendian.yml
@@ -47,7 +47,7 @@ jobs:
           screen -dmS httpserver python3 -m http.server 8000
       - name: Run target OS first time (for setup actions)
         run: >
-          sudo screen -dmS qemu
+          sudo screen -dmS qemu -L -Logfile qemu-console-setup.log
           qemu-system-s390x -M s390-ccw-virtio
           -m 4096 -smp 2 -nographic
           -net nic -net user,hostfwd=tcp::2222-:22
@@ -70,6 +70,20 @@ jobs:
           sleep 5;
           done;
           cat ~/.ssh/known_hosts
+      - name: Wait for disk to be available
+        # virtio_blk is not in the netboot initramfs; it comes from the modloop
+        # image the VM downloads at boot. If that download fails, sshd still
+        # comes up but /dev/vda never appears, so retry the modloop service
+        # until the disk is visible.
+        timeout-minutes: 10
+        run: >
+          while ! ssh -i ssh-key -p2222 root@localhost "test -e /sys/block/vda/device";
+          do
+          echo "Disk not visible, retrying modloop download";
+          ssh -i ssh-key -p2222 root@localhost
+          "rc-service modloop restart; rc-service hwdrivers restart; modprobe virtio_blk 2>/dev/null; ls /lib/modules" || true;
+          sleep 10;
+          done
       - name: Setup alpine to disk
         run: >
           ssh -i ssh-key -p2222 root@localhost "setup-alpine -c setup_alpine_conf &&
@@ -91,7 +105,7 @@ jobs:
           done;
       - name: Run target OS
         run: >
-          sudo screen -dmS qemu
+          sudo screen -dmS qemu -L -Logfile qemu-console.log
           qemu-system-s390x -M s390-ccw-virtio
           -m 4096 -smp 2 -nographic
           -net nic -net user,hostfwd=tcp::2222-:22
@@ -176,3 +190,10 @@ jobs:
             echo "::error::Kernel issues detected in dmesg"
             exit 1
           fi
+      - name: Display QEMU console logs
+        if: ${{ always() }}
+        run: |
+          for f in qemu-console-setup.log qemu-console.log; do
+            echo "===== $f ====="
+            sudo cat "$f" || true
+          done
-- 
cgit v1.2.3