illumos

2018-11-03

01:42 AM illumos gate Bug #9952 (Duplicate): Block size change during zfs receive drops spill block

Replication code in receive_object() falsely assumes that if received object block size is different from local, then... Alexander Motin

2018-08-16

09:42 PM illumos gate Bug #9738: 9112 broke third block copy allocations within one metaslab group

https://github.com/openzfs/openzfs/pull/687 Alexander Motin

01:48 AM illumos gate Bug #9738 (Closed): 9112 broke third block copy allocations within one metaslab group

metaslab_group_alloc_normal() sets tertiary variable for some allocation requests to push those allocations to differ... Alexander Motin

09:20 PM illumos gate Bug #9751: Allocation throttling misplacing ditto blocks

https://github.com/openzfs/openzfs/pull/688 Alexander Motin

09:10 PM illumos gate Bug #9751 (Closed): Allocation throttling misplacing ditto blocks

During synthetic random write benchmarks I found many secondary and tertiary block copies allocations form the same v... Alexander Motin

2018-04-03

06:56 PM illumos gate Bug #9434 (Closed): Speculative prefetch is blocked by device removal code

Testing freshly updated FreeBSD I found that speculative prefetcher does not work for new pools, that never experienc... Alexander Motin

2018-04-02

11:08 PM illumos gate Bug #9433 (Closed): Fix ARC hit rate

https://github.com/zfsonlinux/zfs/pull/6989 needs to be merged. Since dbuf cache size increased recently, it becomes... Alexander Motin

2017-11-20

07:46 PM illumos gate Bug #8835: Speculative prefetch in ZFS not working for misaligned reads

Here is a possible fix: https://github.com/openzfs/openzfs/pull/497 Alexander Motin

05:55 PM illumos gate Bug #8835 (Closed): Speculative prefetch in ZFS not working for misaligned reads

Sequential reads not aligned to block size are not detected by ZFS prefetcher as sequential, killing prefetch and sev... Alexander Motin

2016-12-07

01:38 PM illumos gate Feature #7090: zfs should improve allocation order and throttle allocations

"The queue would be ordered by issued time and offset" is not true, since io_offset field in logical ZIOs is always z... Alexander Motin