Hi, > May be able to test on 6.14.2 on snapshots of the disks tomorrow same issue on 6.15.0-rc3, debug log attached below. The culprit seems to be the call to raid5_recovery_alloc_stripe with noblock set to 0 in r5c_recovery_analyze_meta_block (raid5-cache.c:2155) right after the stripe cache size is raised to 32768. The call to raid5_get_active_stripe unsuccessfully tries to get a free stripe and spawns the reclaim thread and waits for the R5_INACTIVE_BLOCKED flag to be removed (which never happens). With my limited debugging skills it seems like the reclaim thread never does anything in r5l_do_reclaim, because reclaim_target == reclaimable == 0. My attempts to reproduce this without including the member devices (by zero mapping anything but the superblocks) were not successful in triggering the same issue. The relevant part of the journal device is ~12GiB, so uploading the journal is possible. Including (the relevant) parts of the raid members may be more difficult. @Song Liu Sorry for the unsolicited CC, but you seem to be the only main contributor to the raid5-log, that is still frequently active. If there is a better place to issue bug reports like these, please let me know. Linux ubuntu-raid6-recovery 6.15.0-rc3mainline #5 SMP PREEMPT_DYNAMIC Mon Apr 21 18:00:22 CEST 2025 x86_64 x86_64 x86_64 GNU/Linux md: md127 stopped. raid456: run(md127) called. md/raid:md127: device sdf operational as raid disk 0 md/raid:md127: device sdj operational as raid disk 9 md/raid:md127: device sdi operational as raid disk 8 md/raid:md127: device sdg operational as raid disk 7 md/raid:md127: device sdb operational as raid disk 6 md/raid:md127: device sdd operational as raid disk 5 md/raid:md127: device sdk operational as raid disk 4 md/raid:md127: device sde operational as raid disk 3 md/raid:md127: device sda operational as raid disk 2 md/raid:md127: device sdc operational as raid disk 1 md/raid:md127: allocated 10636kB md/raid:md127: raid level 6 active with 10 out of 10 devices, algorithm 2 RAID conf printout: --- level:6 rd:10 wd:10 disk 0, o:1, dev:sdf disk 1, o:1, dev:sdc disk 2, o:1, dev:sda disk 3, o:1, dev:sde disk 4, o:1, dev:sdk disk 5, o:1, dev:sdd disk 6, o:1, dev:sdb disk 7, o:1, dev:sdg disk 8, o:1, dev:sdi disk 9, o:1, dev:sdj md/raid:md127: using device sdh as journal get_stripe, sector 7183404032 __find_stripe, sector 7183404032 __stripe 7183404032 not in cache remove_hash(), stripe 0 init_stripe called, stripe 7183404032 insert_hash(), stripe 7183404032 get_stripe, sector 7183404040 __find_stripe, sector 7183404040 __stripe 7183404040 not in cache remove_hash(), stripe 0 init_stripe called, stripe 7183404040 insert_hash(), stripe 7183404040 get_stripe, sector 7183404048 __find_stripe, sector 7183404048 __stripe 7183404048 not in cache remove_hash(), stripe 0 [..] __find_stripe, sector 7200159744 __stripe 7200159744 not in cache get_stripe, sector 7200159744 __find_stripe, sector 7200159744 __stripe 7200159744 not in cache md/raid:md127: Increasing stripe cache size to 512 to recovery data on journal. +++ raid5d active __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 0 stripes handled --- raid5d inactive +++ raid5d active __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 0 stripes handled --- raid5d inactive get_stripe, sector 7200159744 __find_stripe, sector 7200159744 __stripe 7200159744 not in cache remove_hash(), stripe 0 init_stripe called, stripe 7200159744 insert_hash(), stripe 7200159744 get_stripe, sector 7200159752 __find_stripe, sector 7200159752 __stripe 7200159752 not in cache remove_hash(), stripe 0 init_stripe called, stripe 7200159752 insert_hash(), stripe 7200159752 get_stripe, sector 7200159760 __find_stripe, sector 7200159760 __stripe 7200159760 not in cache get_stripe, sector 7200159760 __find_stripe, sector 7200159760 __stripe 7200159760 not in cache md/raid:md127: Increasing stripe cache size to 1024 to recovery data on journal. +++ raid5d active __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 0 stripes handled --- raid5d inactive +++ raid5d active __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 0 stripes handled --- raid5d inactive +++ raid5d active __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 0 stripes handled --- raid5d inactive +++ raid5d active __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 0 stripes handled --- raid5d inactive +++ raid5d active __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 0 stripes handled --- raid5d inactive +++ raid5d active __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 __get_priority_stripe: handle: empty hold: empty full_writes: 0 bypass_count: 0 0 stripes handled --- raid5d inactive [..] get_stripe, sector 7200204176 __find_stripe, sector 7200204176 __stripe 7200204176 not in cache remove_hash(), stripe 7199919696 init_stripe called, stripe 7200204176 insert_hash(), stripe 7200204176 get_stripe, sector 7200204184 __find_stripe, sector 7200204184 __stripe 7200204184 not in cache remove_hash(), stripe 7199919704 init_stripe called, stripe 7200204184 insert_hash(), stripe 7200204184 get_stripe, sector 7200204192 __find_stripe, sector 7200204192 __stripe 7200204192 not in cache get_stripe, sector 7200204192 __find_stripe, sector 7200204192 __stripe 7200204192 not in cache md/raid:md127: Increasing stripe cache size to 32768 to recovery data on journal. get_stripe, sector 7200204192 __find_stripe, sector 7200204192 __stripe 7200204192 not in cache
