# drbd-9.2.21-rc.2 and drbd-9.3.5-rc.2 **URL:** \ **Category:** Release Announcements\ **Tags:** drbd\ **Created:** [November 11, 2026, 4:32pm UTC](https://forums.linbit.com/t/drbd-9-2-21-rc-2-and-drbd-9-3-5-rc-2/1305 "2026-11-11T16:32:42Z")\ **Posts on this page:** 1\ **Page:** 1
**Author:** ![phil\_reisner](https://avatars.discourse-cdn.com/v4/letter/p/779978/32.png) [@phil\_reisner](https://forums.linbit.com/u/phil_reisner)\ **Post date:** [November 11, 2026, 4:32pm UTC](https://forums.linbit.com/t/drbd-9-2-21-rc-2-and-drbd-9-3-5-rc-2/1305/1 "2026-11-11T16:32:42Z")
Hi, here is the DRBD-9.3.5-rc.2 release. I decided on a second release candidate, as things did not calm down over the last week, which you want before a release. I am now targeting next week for the final release. This is a release candidate. Please test it. 9.3.5-rc.2 (api:genl2/proto:86-101,118-125/transport:23) -------- \* Serve changed block tracking on the drbd netlink family too, so that    it works on kernels without the drbd2 family (before 6.2) \* All fixes from 9.2.21-rc.2   - Fix a sync source running out of memory when the sync target     requests faster than it can read: a target with rs-discard-granularity     keeps its requests within max-buffers, and the source allocates the     pages of a resync read only when it submits the read   - Fix resyncs in opposite directions stopping for good with     max-parallel-resyncs: only the requesting side takes a slot   - Fix an attaching node waiting in WFBitMapT forever while its peer     decided there is nothing to sync   - Fix a failed local write-zeroes leaving the old data on a disk that     stays UpToDate   - Fix auto-promote giving up long before auto-promote-timeout   - Fix disconnect: the peer going to NetworkFailure instead of     Disconnecting, and "drbdadm disconnect" failing with -18 when the     peer tears down first   - Fix TLS: two nodes stalling in the handshake until the TLS timeout,     and going StandAlone when the peer resets the connection during the     handshake   - Fix a use-after-free of the rdma control timer on teardown   - Fix the build on kernels before 4.17 (RHEL 7) and with Linux 7.4,     and the send buffer not growing during a resync on Linux 7.4 9.3.5-rc.1 (api:genl2/proto:86-101,118-125/transport:23) -------- \* Track changed blocks for incremental backups: a tracking bitmap slot    records every write to the backing device; cbt-flush and cbt-start    bracket a snapshot taken while I/O is suspended on all nodes \* suspend-io now also waits for the barrier ack of the last epoch, so    that with protocols A and B the writes are on the peers' stable    storage too; flush-to-peers=no skips that wait \* Add rdma2, a new RDMA transport that moves all data with one-sided    RDMA-WRITE, sends bios zero-copy, and fails over to a surviving path    when a link goes down \* Serve the new drbd2 netlink family next to the drbd family \* Optionally serve DRBD 8.4 userland over the netlink family at version    1 (CONFIG\_DRBD\_NL\_DRBD1, rpmbuild --with nl\_drbd1,    DEB\_BUILD\_OPTIONS=nl\_drbd1); off by default \* Fix lb-tcp skipping a path for good after the peer closed it first    during a disconnect, which starved every later handshake \* Fix spurious and racy "bitmap locked" warnings when bitmap operations    on different peer slots run concurrently \* All fixes from 9.2.21-rc.1   - Fix a resync target staying Inconsistent with nothing left to sync:     when the Primary cannot reach the sync source, after aborting a weak     sync source, and when P\_RS\_CANCEL\_AHEAD overtakes the Ahead state     change   - Fix a reconciliation resync target waiting in WFBitMapT forever, and     out-of-sync bits from before a lost diskless Primary being dropped at     equal dagtags   - Fix a peer's connect-time upgrade from Outdated going unnoticed,     which left writes unreplicated   - Fix an Outdated sync source being marked UpToDate after a resync     with failed blocks   - Fix the day0 slots, a regression in 9.2.20: forget-peer left a slot     that the next UUID bump filled with a later generation, and a peer     created from the day0 UUID could be refused as unrelated data;     attach repairs such meta data   - Fix promoting a peer that holds an older generation of the data a     diskless Primary exposes   - Fix P\_OUT\_OF\_SYNC: never sent for Primaries with node-id 8 or higher,     a crash after del-peer of a Primary, and writes left pending for a     peer that lost its connection   - Fix the activity log: AL updates turned off while a crashed Primary     resyncs, the superblock written too late on "al-updates no", and the     in-core AL not written when AL updates come back on   - Fix a dry-run connect (--tentative) establishing the connection     without a resync   - Fix the last close returning before the peers have the last write on     stable storage, and auto-demote failing on a pending signal   - Fix call-pri-lost-after-sb killing the receiver instead of demoting   - Fix online resize: drain application I/O before the commit and before     a participant replies, cache the peers' size answers per volume, and     refuse a too small peer while syncing from a larger one   - Fix a receiver deadlock at the end of a resync when the link fails     while the bitmap is written, and bound the memory a bitmap write-out     holds   - Fix resync pacing: the fixed resync-rate (c-plan-ahead 0) holds, the     resync queue at the backing device is bounded so application I/O does     not wait behind it, and DRBD logs when max-buffers or c-max-rate     limit a resync   - Fix connect and disconnect issues: a spurious NetworkFailure right     after a handshake, a stale ping timeout dropping the next connect, a     peer's close while connecting taken as a network error, ko-count 0     being counted down, and a second peer accepted beside a DRBD 8.4 peer   - Fix lb-tcp: a NULL pointer dereference in the receiver, a CPU spin     when the peer closes mid-payload, zero-length data chunks, hot-adding     paths, and listeners left on unused paths   - Fix rdma: use-after-free and refcount underflows on teardown during     connect, a 60 s hang after a failed establish, going StandAlone on an     interrupted connect, and a crash with a peer running rdma2   - Fix the TLS handshake compat code (upstream fixes): leaks, refcount     and use-after-free bugs, and a missing admin check on DONE   - Fix memory corruption by peer\_device flags on 32-bit kernels   - Fix a negative ap-in-flight after a resend, BarrierAck count warnings,     sending on after a send failed, and an unsent packet flushed at     disconnect ("Wrong magic value" on the peer)   - Fix attach hanging after the node was a verify target, and a double     release when linking the holder fails   - Show the submit and sent ages of requests in debugfs again, and mark     peer writes that completed locally   - Fix building SUSE KMPs on SLES 16, and the build with     CONFIG\_DRBD\_TIMING\_STATS [https://pkg.linbit.com//downloads/drbd/9/drbd-9.3.5-rc.2.tar.gz](https://pkg.linbit.com//downloads/drbd/9/drbd-9.3.5-rc.2.tar.gz) > > > Signed-off-by: Philipp Reisner \ [https://pkg.linbit.com//downloads/drbd/9/drbd-9.2.21-rc.2.tar.gz](https://pkg.linbit.com//downloads/drbd/9/drbd-9.2.21-rc.2.tar.gz) > > > Signed-off-by: Philipp Reisner \ cheers, Phil