Known Issues in Lightbits 3.18.1

48930

In a rare situation where during a rebuild at least two snapshots are deleted, one containing data and a higher one containing a trim to the same blocks, the trim might be applied to the next existing version/snapshot - thus losing the valid data in it.

48532

The node-manager service can fail to start if the number of drives present falls below the minimum required for the node's configured layout, for example, after a drive failure, power loss or planned maintenance. In earlier releases, the node started with the missing drive marked as failed; in affected releases it does not start at all. Workaround: Set initialDeviceCount in system-profile.yaml to a lower value. This issue is fixed in v3.18.4, v3.20.2, and v3.21.1.

48435

On some AMD EPYC servers, NVMe SSDs can report I/O timeouts. This can cause a storage node to be marked as failed, and while the node is out of service, volumes with reduced protection can become read-only. Workaround: Disable PCIe Relaxed Ordering in the server BIOS.

48205

When a large number of clone volumes are created from a small number of shared snapshots and a node is then restarted or replaced, some of those clone volumes can become stuck retrying volume creation, which can leave the node's protection groups unable to complete their rebuild. This issue is fixed in v3.20.2 and v3.21.1.

48041

In affected versions, when a management operation fails repeatedly, the delay between retries eventually drops to almost zero: after about 1.5 minutes of consecutive failures in v3.18.x, or about 16 hours in v3.19.x and v3.20.x. The continuous retries can keep a CPU core busy and flood the logs, pushing out older log entries needed to investigate the original failure. This issue is fixed in v3.18.4, v3.20.2, and v3.21.1.

47753

Changing a node's Permanent Failure (PF) threshold can behave unexpectedly if the node is already Inactive. Lowering the threshold to a value that has not yet elapsed does not take effect until the original threshold time passes, and raising the threshold can prevent the node from reaching Permanent Failure at all. Workaround: Change a node's PF threshold before maintenance that will make it inactive. To shorten the threshold of a node that is already inactive, use a value that has already elapsed.

47343

In rare cases, deleting a snapshot while a rebuild is in progress can raise DataIntegrityDuringRebuild events, and the affected data is not rebuilt to the secondary replica. The frontend log on the primary shows "Rebuild read detected a damaged object". This issue is fixed in v3.18.3, v3.20.2, and v3.21.1.

47300

If the cluster manager's connection to the cluster's internal coordination service is interrupted (for example, due to a leader change, network disruption, or a brief service restart), the cluster manager's view of cluster servers can stop updating entirely, with no automatic recovery. Workaround: Restart the cluster-manager service to restore normal operation.

46209

The bundled Grafana Alloy log-streaming agent ships with Grafana's anonymous usage statistics reporting enabled by default. The report is sent out on a fixed four-hour schedule to https://stats.grafana.org/. In air-gapped or firewalled deployments, these attempts fail and produce recurring error entries in the journal. In connected deployments, they could represent an unsanctioned outbound connection to a third party. The exact list of fields collected is documented at https://grafana.com/docs/alloy/latest/data-collection/.

45729

When SSD journaling is configured with multiple NVMe devices, cluster cleanup does not remove the RAID (mdadm) arrays created by Lightbits. Before re-deploying on the same servers, manually remove these arrays using mdadm.

45712

In rare cases, if the internal connection used to track volume or snapshot state changes is interrupted (for example, due to a leader change, network interruption, or a brief service restart), a volume or snapshot update can remain stuck showing as still in progress even though the underlying change has completed. Workaround: Restart node-manager on the affected server to resolve the stuck state.

45625

After running the Ansible node cleanup, leftover journaling RAID (mdadm) devices can keep drives attached, so the next installation detects fewer NVMe drives than expected. Workaround: Stop the mdadm RAID device manually before reinstalling. This issue is fixed in v3.21.1.

45202

When a volume is created with lbcli, any part of a label key before a hyphen is dropped. For example, --labels="pve-vmid=100" is saved with the key "vmid" instead of "pve-vmid". Workaround: Avoid hyphens in label keys, or create the volume through the REST API, which is not affected.

45157

During a node power-up after an abrupt shutdown, the node's reported power-up progress can reach 100%, drop back to 50-60%, and then climb to 100% again. This affects the reported percentage only. Recovery completes correctly, and the node returns to service only when it is actually done. No workaround is needed.

44875

lbcli and REST documents describe that you can evict a server with 1x volumes with the force flag, but this is not possible. It also indicates that you can evict a server even if it could lead to risk of service loss using the force flag. However, force will only allow you to evict a server if its nodes are unavailable but still requires other nodes to be active.

44673

In rare cases, performing a KEK rotation while simultaneously executing a high volume of volume/snapshot control plane operations could result in increased etcd contention. This affects only customers utilizing cluster-level encryption with the KEK rotation API.

44464

In affected versions, if the server selected as the new primary for an in-progress volume migration becomes permanently failed, the migration does not abort and instead retries indefinitely, repeatedly updating the cluster state. This issue is fixed in v3.21.1.

44435

pmem_init and lbe executables are linked using staticx and therefore require /tmp to be mounted with the exec option enabled; otherwise, these executables will fail to run with "permission denied".

44383

When cluster-level encryption is enabled, deleting a server that hosts the active Cluster Manager (CM) could prevent subsequent server additions from completing successfully. After removing such a server from the cluster, stop the CM service on it or shut the server down.

44334

In affected versions, stopping or restarting the node-manager service on a server (for example, during maintenance or an upgrade) can leave the data-path service hung while it waits for a configuration update that never arrives, which delays or blocks the server's shutdown. This issue is fixed in v3.18.3 and v3.21.1.

44298

If a journal NVMe device fails when the node manager is down, then when the node manager comes back up, it can try to take another disk to use for journaling. If the matchers for journaling are the same as for data devices, it can take one of the data devices by mistake - causing the GFTL to crash. For journaling, it is necessary to use specific matchers that are different then the data devices (for example, the serial number).

43456

In a rare combination of conditions, a storage node could fail to start after a restart if placement group membership changed while the node was offline. This requires memory pressure during a prior recovery (causing stale metadata to be persisted), followed by placement group rebalancing that removes volumes from the node. Normal operations and graceful recovery flows are not affected.

42991

In a rare case, a network disconnect coinciding precisely with an NVMe SSD device state change could prevent subsequent state updates for that specific device, so changes to its health might not be reflected. The condition resolves automatically the next time the Node Manager service is restarted.

42963

When SSD Journaling is enabled, if the frontend fails due to a non-journal-related issue, the Node Manager (NM) might incorrectly classify the failure as a journal device failure, causing the NM to remain inactive and enter a Permanently Failed state. When SSD journaling is disabled, a frontend failure does not impact the failure scenario. However, users might receive a spurious "journal device failed" event even when journaling is not in use. This is cosmetic only and does not reflect an actual journal issue.

42800

A volume's protections state could fail to update correctly in some cases of network/ETCD unavailability.

42614

Cluster manager and etcd services could suffer a very slow potential memory leak in rare cases. Mishandling of a deprecated GFTL data loss event could cause the event clean logic to stop cleanup of old events, leading to a continuous increase in the number of events stored on the cluster.

42309

In rare cases, a CM failover occurring during the initial KEK rotation process may result in a race condition where the new CM fails to become active, causing some API calls to fail. The data path remains unaffected as long as all nodes are healthy.

41873

Under a specific race condition, if a snapshot is created while a node is down and subsequently deleted during a very precise window in the node's startup sequence, the node may become unavailable.

41466

Creating a snapshot with a retention time greater than 192 years will fail and cause the API service to restart.

41095

The NodeRebuildNotPossible alert may not trigger under conditions where it should, resulting in missed notifications for rebuild-blocking scenarios.

40883

When using the vCenter Plugin (VCP) to upgrade a cluster to Lightbits 3.17.1 or later, the upgrade fails because VCP cannot parse the updated version format. To complete the upgrade, use the Lightbits core CLI or REST API directly.

40428

In extremely rare cases, the reported logical size of a volume could be incorrect after a discard operation is performed and TRIM support was enabled.

38706

In some rare cases, the frontend could hang during shutdown.

38497

When creating a new server to replace another server in the cluster using the --extend-cluster=false flag (which is the default setting), and at a much later time this server and its node experience a permanent failure and fail in place is enabled (causing the servers resources to be migrated), if the server goes active again it might not participate in all proper distribution of replicas over the cluster and could cause an imbalance of resources.

37830

In extremely rare cases, a node may not recover to an active state if an I/O error or bad block is encountered on an underlying SSD during its startup sequence. This prevents a key service (gftl) from initializing correctly and may require manual intervention (such as removing the failed SSD from the system) to allow the node to complete its recovery successfully.

37505

In a rare combination of events, the physicalOwnedCapacity volume statistic may report an incorrect value when data at a specific LBA is overwritten with content that has a different compression ratio. This affects the reported statistic only; stored data and I/O are unaffected.

33865

In certain cases when migrating volumes during dynamic rebalancing, a VolumeInDegradedProtectionState event could be sent out when the volume is actually fully protected.

28027

In specific circumstances, the server upgrade status may not reflect the correct state after a successful retry. This occurs in the following sequence: 1. A server is upgraded to version x.y.z. 2. The operation fails (for example, due to a timeout); however, the server binaries are successfully updated to version x.y.z. 3. The upgrade is retried to version x.y.z — this step is skipped internally, as the binaries are already up to date. 4. The upgrade status continues to show the previous failed operation, even though the retry completed without error.