The TrueNAS team is pleased to release TrueNAS 25.10.3!
Notable changes:
- Fixes read authentication errors caused by block cloning between encrypted ZVOLs with different encryption keys (NAS-140263). Block cloning between encrypted ZVOLs that did not share compatible encryption keys was allowed, but subsequent reads of the cloned data produced authentication failures. TrueNAS now verifies that encryption keys match before allowing a block clone operation to proceed.
- Fixes Active Directory users and groups not found when using the idmap_ad backend with RFC2307 (NAS-140077). The idmap_ad backend failed when smb.conf was configured to disallow a private krb5.conf, preventing Active Directory users and groups from being resolved. Shares and services that depend on directory service identities were inaccessible on affected systems.
- Fixes pool creation errors on systems with a USB boot drive or other disk with a duplicate serial number (NAS-139947). The disk availability check incorrectly blocked pool creation if any disk on the system had a duplicate serial number, even when that disk was unrelated to the new pool. USB boot drives commonly have blank or duplicate serial numbers and were the most frequent cause. Note that TrueNAS does not recommend USB boot devices. Installing on SATA, SAS, or NVMe flash media is recommended.
- Fixes issues with log VDEV removal (NAS-140080). Resolves errors that could occur when removing a log VDEV from a storage pool. Affected systems encountered failures during the log VDEV removal process.
- Fixes link aggregation failover not working out of the box (NAS-136978). Resolves an issue where the MII polling interval was not correctly configured for bonding interfaces using the FAILOVER protocol. The bonding driver could not monitor slave link status, causing failover to fail silently when an interface lost carrier.
- Fixes Netlink socket errors that caused network interface operations to fail (NAS-140041). Concurrent netlink route operations could collide at the kernel level, producing “Netlink socket busy” (EBUSY) errors. This was most visible as bond status alert failures. The fix serializes netlink route operations to prevent collisions, increases the retry limit, and adds a short delay between retries to reduce the chance of repeated failures.
- Fixes NTB transport memory allocation failures during link state changes (NAS-139627). Non-Transparent Bridge (NTB) transport memory windows were allocated and freed on every link up/down event. Allocating several megabytes of physically contiguous memory on a running system could fail, causing instability. Memory windows are now pre-allocated at boot, avoiding repeated allocation attempts during link changes. This primarily affects TrueNAS Enterprise High Availability (HA) systems.
- Fixes stale NFS file handles for ZFS snapshots that could not be recovered without a client reboot (NAS-140474). When mountd failed to resolve a snapshot’s filesystem ID — such as after failover, snapshot expiration, or an automount failure — it permanently cached a negative response, causing all subsequent file handle lookups to return ESTALE errors. Client-side remounts did not clear the cached state. The fix replaces the permanent negative cache with a 120-second retry window, allowing recovery once the snapshot becomes accessible again.
- Documentation improvement: Adds Drive Health Management documentation covering how TrueNAS 25.10 monitors drive health through ZFS real-time failure detection, automated 90-minute SMART polling, and improved alert filtering. Also covers manual SMART test scheduling options available to TrueNAS Community Edition users. See the Drive Health Management tutorial for full details.
See the Release Notes and changelog for more details.
25.10.3 Documentation : https://www.truenas.com/docs/scale/25.10
Download : https://www.truenas.com/download-truenas-community-edition
Thank you for using TrueNAS! As always, we appreciate your feedback!
