1 #+BIBLIOGRAPHY: refs.bib
5 BTRFS is a Linux filesystem based on copy-on-write, allowing for
6 efficient snapshots and clones.
8 It uses B-trees as its main on-disk data structure. The design goal is
9 to work well for many use cases and workloads. To this end, much
10 effort has been directed to maintaining even performance as the
11 filesystem ages, rather than trying to support a particular narrow
14 Linux filesystems are installed on smartphones as well as enterprise
15 servers. This entails challenges on many different fronts.
17 - Scalability :: The filesystem must scale in many dimensions: disk
18 space, memory, and CPUs.
20 - Data integrity :: Losing data is not an option, and much effort is
21 expended to safeguard the content. This includes checksums, metadata
22 duplication, and RAID support built into the filesystem.
24 - Disk diversity :: The system should work well with SSDs and hard
25 disks. It is also expected to be able to use an array of different
26 sized disks, which poses challenges to the RAID and striping
30 *** [2023-08-08 Tue] btrfs performance speculation :: 31 - [[https://www.percona.com/blog/taking-a-look-at-btrfs-for-mysql/]] 32 - zfs outperforms immensely, but potential misconfiguration on btrfs side (virt+cow 34 - https://www.ctrl.blog/entry/btrfs-vs-ext4-performance.html 35 - see the follow up comment on this post 36 - https://www.reddit.com/r/archlinux/comments/o2gc42/is_the_performance_hit_of_btrfs_serious_is_it/ 38 I’m the author of OP’s first link. I use BtrFS today. I often shift lots of 39 de-duplicatable data around, and benefit greatly from file cloning. The data is actually 40 the same data that caused the slow performance in the article. BtrFS and file cloning 41 now performs this task quicker than a traditional file system. (Hm. It’s time for a 44 In a laptop with one drive: it doesn’t matter too much unless you do work that benefit 45 from file cloning or snapshots. This will likely require you to adjust your tooling and 46 workflow. I’ve had to rewrite the software I use every day to make it take advantage of 47 the capabilities of a more modern file system. You won’t benefit much from the data 48 recovery and redundancy features unless you’ve got two storage drives in your laptop and 49 can setup redundant data copies. 51 on similar hardware to mine? 53 It’s not a question about your hardware as much as how you use it. The bad performance I 54 documented was related to lots and lots of simultaneous random reads and writes. This 55 might not be representative of how you use your computer. 57 - https://dl.acm.org/doi/fullHtml/10.1145/3386362 58 - this is about distributed file systems (in this case Ceph) - they argue against 59 basing DFS on ondisk-format filesystems (XFS ext4) - developed BlueStore as 60 backend, which runs directly on raw storage hardware. 61 - this is a good approach, but expensive (2 years in development) and risky 62 - better approach is to take advantage of a powerful enough existing ondisk-FS 63 format and pair it with supporting modules which abstract away the 'distributed' 65 - the strategy presented here is critical for enterprise-grade hardware where the 66 ondisk filesystem becomes the bottleneck that you're looking to optimize 67 - https://lore.kernel.org/lkml/cover.1676908729.git.dsterba@suse.com/ 68 - linux 6.3 patch by David Sterba [2023-02-20 Mon] 69 - btrfs continues to show improvements in the linux kernel, ironing out the kinks 70 - makes it hard to compare benchmarks tho :/ 72 - see this WIP k-ext for macos: [[https://github.com/relalis/macos-btrfs][macos-btrfs]] 73 - maybe we can help out with the VFS/mount support
75 - [[https://btrfs.readthedocs.io/en/latest/dev/On-disk-format.html][on-disk-format]] 76 - 'btrfs consists entirely of several trees. the trees use copy-on-write.'
77 - trees are stored in nodes which belong to a level in the b-tree structure.
78 - internal nodes (inodes) contain refs to other inodes on the
/next/ level OR
79 - to leaf nodes then the level reaches 0.
80 - leaf nodes contain various types depending on the tree.
82 - 0:8 uint
= objectid, each tree has its own set of object IDs 83 - 8:1 uint = item type
84 - 9:8 uint
= offset, depends on type. 88 - primary superblock is located at 0x10000 (64KiB) 89 - Mirror copies of the superblock are located at physical addresses 0x4000000 (64 90 MiB) and 0x4000000000 (256GiB), if valid. copies are updated simultaneously. 91 - during mount only the first super block at 0x10000 is read, error causes mount to 93 - BTRFS onls recognizes disks with a valid 0x10000 superblock. 95 - stored at the start of every inode 96 - data following it depends on whether it is an internal or leaf node. 98 - node header followed by a number of key pointers 100 - 11:8 uint = block number
101 - 19:8 uint
= generation 103 - leaf nodes contain header followed by key pointers 105 - 11:4 uint = data offset relative to end of header(65)
106 - 15:4 uint
= data size 109 - holds ROOT_ITEMs, ROOT_REFs, and ROOT_BACKREFs for every tree other than itself. 110 - used to find the other trees and to determine the subvol structure. 111 - holds items for the 'root tree directory'. laddr is store in the superblock 113 - free ids: BTRFS_FIRST_FREE_OBJECTID=256ULL:BTRFS_LAST_FREE_OBJECTID=-256ULL
114 - otherwise used for internal use
115 *** send-stream format 116 - [[https://btrfs.readthedocs.io/en/latest/dev/dev-send-stream.html][send stream format]] 117 - Send stream format represents a linear sequence of commands describing actions to be
118 performed on the target filesystem (receive side), created on the source filesystem
120 - The stream is currently used in two ways: to generate a stream representing a
121 standalone subvolume (full mode) or a difference between two snapshots of the same
122 subvolume (incremental mode).
123 - The stream can be generated using a set of other subvolumes to look for extent
124 references that could lead to a more efficient stream by transferring only the
125 references and not full data.
126 - The stream format is abstracted from on-disk structures (though it may share some
127 BTRFS specifics), the stream instructions could be generated by other means than the
129 - it's a checksum+TLV
130 - header: u32len,u16cmd,u32crc32c
131 - data: type,length,raw data
132 - the v2 protocol supports the encoded commands
133 - the commands are kinda clunky - need to MKFIL/MKDIR then RENAM to create
134 *** [2023-08-09 Wed] ioctls 136 - https://docs.kernel.org/userspace-api/ioctl/ioctl-number.html
137 - Btrfs filesystem some lifted to vfs/generic
138 - fs/btrfs/ioctl.h and linux/fs.h
142 - core component of TrueNAS software
150 -- [cite/t/f:@xfs-scalability]
153 -- [cite/t/f:@hd-failure-ml]
155 -- [cite/t/f:@smart-ssd-qp]
156 -- [cite/t/f:@ssd-perf-opt]
159 -- [cite/t/f:@flash-openssd-systems]
161 -- [cite/t/f:@nvme-ssd-ux]
162 --
[[https://nvmexpress.org/specifications/][specifications]] 164 -- [cite/t/f:@zns-usenix]
166 Zoned Storage is an open source, standards-based initiative to enable data centers to
167 scale efficiently for the zettabyte storage capacity era. There are two technologies
168 behind Zoned Storage, Shingled Magnetic Recording (SMR) in ATA/SCSI HDDs and Zoned
169 Namespaces (ZNS) in NVMe SSDs.
171 --
[[https://zonedstorage.io/][zonedstorage.io]] 172 -- $465 8tb 2.5"?
[[https://www.serversupply.com/SSD/PCI-E/7.68TB/WESTERN%20DIGITAL/WUS4BB076D7P3E3_332270.htm][retail]] 174 -- [cite/t/f:@emmc-mobile-io]
178 - [[https://elixir.bootlin.com/linux/latest/source/Documentation/userspace-api/ioctl/ioctl-number.rst][ioctl-numbers]] 182 - [[https://crates.io/crates/nix][crates.io]] 184 - [[https://crates.io/crates/memmap2][crates.io]] 186 - [[https://crates.io/crates/zstd][crates.io]] 188 - [[https://crates.io/crates/rocksdb][crates.io]] 190 - [[https://crates.io/crates/tokio][crates.io]] 192 - [[https://crates.io/crates/tracing][crates.io]] 193 **** tracing-subscriber 194 - [[https://crates.io/crates/tracing-subscriber][crates.io]] 196 - [[https://crates.io/crates/axum][crates.io]] 198 - [[https://crates.io/crates/tower][crates.io]] 200 - [[https://crates.io/crates/uuid][crates.io]] 203 - [[https://github.com/rust-lang/rust/issues/109736][tracking-issue]] 204 *** {BTreeMap,BTreeSet}::extract_if 205 - [[https://github.com/rust-lang/rust/issues/70530][tracking-issue]] 208 - [[https://gitlab.common-lisp.net/asdf/asdf][gitlab.common-lisp.net]] 209 - [[https://asdf.common-lisp.dev/][common-lisp.dev]] 210 - [[https://github.com/fare/asdf/blob/master/doc/best_practices.md][best-practices]] 212 ** Reference Projects 214 - [[https://github.com/stumpwm/stumpwm][github]] 216 - [[https://github.com/atlas-engineer/nyxt][github]] 218 - [[https://github.com/kaveh808/kons-9][github]] 220 - [[https://github.com/vindarel/cl-torrents][github]] 222 - [[https://github.com/froggey/Mezzano][github]] 224 - [[https://github.com/whily/yalo][github]] 226 - [[https://github.com/ledger/cl-ledger][github]] 228 - [[https://github.com/lem-project/lem][github]] 230 - [[https://github.com/kindista/kindista][github]] 232 - [[https://github.com/ryukinix/lisp-chat][github]] 234 #+print_bibliography:
