2026-11-16 - 2026-11-23
Overview
25 Pull requests merged by 2 users
Merged
#114 Store v1: Fix two flusher bugs with deletions
Merged
#137 vitastor-disk: Exit with 1 if prepare fails to initialize an OSD, list skipped disks
Merged
#152 Report invalid osd_network as a fatal error instead of an exception
Merged
#149 Store v1: Don't let the ack of an older big write unblock later small writes
Merged
#146 osd: Use one timestamp for picking recovery and arming its retry timer
Merged
#120 Never fall back to /dev/nbd0 in vitastor-nbd unmap and revive
Merged
#134 vitastor-disk: Accept symlinks to whole disks in prepare
Merged
#143 osd: Pause recovery of a PG after an error instead of retrying it in a loop
Merged
#142 Fix vitastor-nbd spinning forever after NBD_CMD_DISC
Merged
#136 Store v2: Do not return unstable entries beyond the listing page
Merged
#133 cli/dd: Keep in-order output going after ignored errors
Merged
#135 vitastor-disk: Refuse a disk given both as data and in --fast-devices
Merged
#141 osd: Stop on SIGINT/SIGTERM from the main loop, not the signal handler
Merged
#121 Report netlink-map failures in its exit status
Merged
#138 vitastor-disk: pass config_path to vitastor-cli/vitastor-osd
Merged
#128 Store v1: Do not skip checksum verification of reads that start after a parallel flush
Merged
#132 osd: Test peering flush batches waiting for the PG history
Merged
#129 Store v1: Don't let a sync of an older big write unblock later small writes
Merged
#131 osd: Keep the holders of unstable EC versions in the PG history and accept them from pg_data_size roles
Merged
#126 vitastor-disk: Write csum_block_size 0 in JE_START without data checksums
Merged
#125 allocator: Fix find_free() returning a block past the end when full
Merged
#127 vitastor-disk: Fix stable/rollback handling in index_journal_by_object
Merged
#119 Free NBD netlink sockets with nl_socket_free()
Merged
#118 osd: Recalculate object state online when scrub finds its target version deleted
Merged
#117 Client: Retry a failed config reload instead of failing the running client
12 Pull requests proposed by 2 users
Proposed
#122 client: resend only the unsent tail after a short io_uring send
Proposed
#123 WIP: [DNM] osd: run calc_object_states off the OSD thread
Proposed
#130 cli/dd: Write an O_APPEND stdout in order
Proposed
#139 vitastor-disk: Check prepare options before partitioning, don't reuse damaged OSD partitions
Proposed
#140 osd: Wait for the old state lease to expire instead of exiting
Proposed
#144 Subscribe to EPOLLOUT only while a send waits for it in no-io_uring mode
Proposed
#145 Test dd conv=noerror with a failing in-order write
Proposed
#147 osd: Test a rewrite after a delete that missed a down OSD
Proposed
#150 osd: Log only the first ENOSPC write subop failure per slow_log_interval
Proposed
#151 Capture client init errors in qemu_driver without io_uring too
Proposed
#153 Add client_exit_on_error and client_infinite_start docs
Proposed
#155 Refuse to revive an NBD device with another image
3 Unresolved Conversations
Open
#107
Store v2: Wait for fsyncs submitted by other operations before completing a sync
Open
#106
osd: Roll back unstable writes of deleted pools in the old store
Open
#115
WIP: osd: Stabilize unstable writes of a removed pool before stopping its PGs