From 8151e1e41a9b2b7e28ce128fdd6c258c6453a505 Mon Sep 17 00:00:00 2001 From: Andrei Kvapil Date: Wed, 7 Jan 2026 13:27:06 +0100 Subject: [PATCH] fix(linstor): prevent orphaned DRBD devices during toggle-disk retry The previous retry logic in toggle-disk removed layer data from controller DB and recreated it. However, removeLayerData() only deletes from the database without calling drbdadm down on the satellite, leaving orphaned DRBD devices in the kernel that occupy ports and block new operations. This fix changes retry to simply repeat the operation with existing layer data, allowing the satellite to handle it idempotently. Upstream: https://github.com/LINBIT/linstor-server/pull/475 Co-Authored-By: Claude Signed-off-by: Andrei Kvapil --- .../patches/allow-toggle-disk-retry.diff | 26 +++++++------------ 1 file changed, 10 insertions(+), 16 deletions(-) diff --git a/packages/system/linstor/images/piraeus-server/patches/allow-toggle-disk-retry.diff b/packages/system/linstor/images/piraeus-server/patches/allow-toggle-disk-retry.diff index 4f4f2fc6..264e0221 100644 --- a/packages/system/linstor/images/piraeus-server/patches/allow-toggle-disk-retry.diff +++ b/packages/system/linstor/images/piraeus-server/patches/allow-toggle-disk-retry.diff @@ -1,8 +1,8 @@ diff --git a/controller/src/main/java/com/linbit/linstor/core/apicallhandler/controller/CtrlRscToggleDiskApiCallHandler.java b/controller/src/main/java/com/linbit/linstor/core/apicallhandler/controller/CtrlRscToggleDiskApiCallHandler.java -index 1a6f7b7f0..bd447e049 100644 +index d93a18014..cc8ce4f04 100644 --- a/controller/src/main/java/com/linbit/linstor/core/apicallhandler/controller/CtrlRscToggleDiskApiCallHandler.java +++ b/controller/src/main/java/com/linbit/linstor/core/apicallhandler/controller/CtrlRscToggleDiskApiCallHandler.java -@@ -58,7 +58,9 @@ import com.linbit.linstor.stateflags.StateFlags; +@@ -57,7 +57,9 @@ import com.linbit.linstor.stateflags.StateFlags; import com.linbit.linstor.storage.StorageException; import com.linbit.linstor.storage.data.adapter.drbd.DrbdRscData; import com.linbit.linstor.storage.interfaces.categories.resource.AbsRscLayerObject; @@ -12,7 +12,7 @@ index 1a6f7b7f0..bd447e049 100644 import com.linbit.linstor.storage.utils.LayerUtils; import com.linbit.linstor.tasks.AutoDiskfulTask; import com.linbit.linstor.utils.layer.LayerRscUtils; -@@ -317,21 +319,90 @@ public class CtrlRscToggleDiskApiCallHandler implements CtrlSatelliteConnectionL +@@ -387,21 +389,84 @@ public class CtrlRscToggleDiskApiCallHandler implements CtrlSatelliteConnectionL Resource rsc = ctrlApiDataLoader.loadRsc(nodeName, rscName, true); @@ -61,18 +61,12 @@ index 1a6f7b7f0..bd447e049 100644 + .concatWith(ctrlRscDfnApiCallHandler.get().updateProps(rsc.getResourceDefinition())); + } + // If adding disk and DISK_ADD_REQUESTED is already set, treat as retry -+ // First clean up partially created storage by removing and recreating layer data ++ // Simply retry the operation with existing layer data - satellite will handle it idempotently ++ // NOTE: We don't remove/recreate layer data here because removeLayerData() only deletes ++ // from controller DB without calling drbdadm down on satellite, leaving orphaned DRBD devices + errorReporter.logInfo( -+ "Toggle Disk retry on %s/%s - DISK_ADD_REQUESTED already set, cleaning up and retrying", ++ "Toggle Disk retry on %s/%s - DISK_ADD_REQUESTED already set, retrying operation", + nodeNameStr, rscNameStr); -+ -+ // Remove old layer data and recreate to ensure clean state -+ // This forces satellite to delete any partially created storage and start fresh -+ LayerPayload payload = new LayerPayload(); -+ copyDrbdNodeIdIfExists(rsc, payload); -+ List layerList = removeLayerData(rsc); -+ ctrlLayerStackHelper.ensureStackDataExists(rsc, layerList, payload); -+ + ctrlTransactionHelper.commit(); + return Flux + .just(new ApiCallRcImpl()) @@ -113,7 +107,7 @@ index 1a6f7b7f0..bd447e049 100644 } if (!removeDisk && !ctrlVlmCrtApiHelper.isDiskless(rsc)) -@@ -342,17 +413,43 @@ public class CtrlRscToggleDiskApiCallHandler implements CtrlSatelliteConnectionL +@@ -412,17 +477,43 @@ public class CtrlRscToggleDiskApiCallHandler implements CtrlSatelliteConnectionL true )); } @@ -160,7 +154,7 @@ index 1a6f7b7f0..bd447e049 100644 if (removeDisk) { // Prevent removal of the last disk -@@ -1324,6 +1421,30 @@ public class CtrlRscToggleDiskApiCallHandler implements CtrlSatelliteConnectionL +@@ -1446,6 +1537,30 @@ public class CtrlRscToggleDiskApiCallHandler implements CtrlSatelliteConnectionL } } @@ -191,7 +185,7 @@ index 1a6f7b7f0..bd447e049 100644 private void markDiskAdded(Resource rscData) { try -@@ -1389,6 +1510,41 @@ public class CtrlRscToggleDiskApiCallHandler implements CtrlSatelliteConnectionL +@@ -1511,6 +1626,41 @@ public class CtrlRscToggleDiskApiCallHandler implements CtrlSatelliteConnectionL return layerData; }