X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=Documentation%2Fpublic-inbox-tuning.pod;h=53668eccb7cb85a7a7c0f6149c0c8c54189614e2;hb=d0bc58517c97c8982fdf42533297d96be9efb484;hp=abc53d1ee0722c7561190bc805dcf5b4e6b81993;hpb=99850fabd5fc628ab29c718e9d7de09b8114b208;p=public-inbox.git diff --git a/Documentation/public-inbox-tuning.pod b/Documentation/public-inbox-tuning.pod index abc53d1e..53668ecc 100644 --- a/Documentation/public-inbox-tuning.pod +++ b/Documentation/public-inbox-tuning.pod @@ -16,7 +16,7 @@ New inboxes: public-inbox-init -V2 =item 2 -Process spawning +Optional Inline::C use =item 3 @@ -34,6 +34,14 @@ Performance on solid state drives Read-only daemons +=item 7 + +Other OS tuning knobs + +=item 8 + +Scalability to many inboxes + =back =head2 New inboxes: public-inbox-init -V2 @@ -42,7 +50,7 @@ If you're starting a new inbox (and not mirroring an existing one), the L<-V2|public-inbox-v2-format(5)> requires L, but is orders of magnitude more scalable than the original C<-V1> format. -=head2 Process spawning +=head2 Optional Inline::C use Our optional use of L speeds up subprocess spawning from large daemon processes. @@ -52,9 +60,17 @@ environment variable to point to a writable directory, or create C<~/.cache/public-inbox/inline-c> for any user(s) running public-inbox processes. +If libgit2 development files are installed and L +is enabled (described above), per-inbox C +processes are replaced with a single L process running +C in read-only daemons. libgit2 use +will be available in public-inbox 1.7.0+ + More (optional) L use will be introduced in the future to lower memory use and improve scalability. +Note: L is required for L, but not public-inbox-* + =head2 Performance on rotational hard disk drives Random I/O performance is poor on rotational HDDs. Xapian indexing @@ -69,11 +85,12 @@ footprint when indexing on HDDs. Initializing a mirror with a high C<--jobs> count to create more shards (in C<-V2> inboxes) will keep each shard smaller and -reduce its kernel page cache footprint. +reduce its kernel page cache footprint. Keep in mind excessive +sharding imposes a performance penalty for read-only queries. Users with large amounts of RAM are advised to set a large value for C as documented in -L. +L. C users on Linux 4.0+ are advised to try the C<--perf-same_cpu_crypt> C<--perf-submit_from_crypt_cpus> @@ -88,12 +105,23 @@ used by public-inbox are no exception to that. public-inbox 1.6.0+ disables copy-on-write (CoW) on Xapian and SQLite indices on btrfs to achieve acceptable performance (even on SSD). -Disabling copy-on-write also disables checksumming, thus raid1 -(or higher) configurations may corrupt on unsafe shutdowns. +Disabling copy-on-write also disables checksumming, thus C +(or higher) configurations may be corrupt after unsafe shutdowns. Fortunately, these SQLite and Xapian indices are designed to recoverable from git if missing. +Disabling CoW does not prevent all fragmentation. Large values +of C also limit fragmentation during +the initial index. + +Avoid snapshotting subvolumes containing Xapian and/or SQLite indices. +Snapshots use CoW despite our efforts to disable it, resulting +in fragmentation. + +L can be used to monitor fragmentation, and +C may be necessary. + Large filesystems benefit significantly from the C mount option documented in L. @@ -106,6 +134,11 @@ While SSD read performance is generally good, SSD write performance degrades as the drive ages and/or gets full. Issuing C commands via L or similar is required to sustain write performance. +Users of the Flash-Friendly File System +L may benefit from +optimizations found in SQLite 3.21.0+. Benchmarks are greatly +appreciated. + =head2 Read-only daemons L, L, and @@ -116,12 +149,32 @@ for parallelism. The open file descriptor limit (C, C in L, C in L) may need to be raised to -accomodate many concurrent clients. +accommodate many concurrent clients. Transport Layer Security (IMAPS, NNTPS, or via STARTTLS) significantly increases memory use of client sockets, sure to account for that in capacity planning. +=head2 Other OS tuning knobs + +Linux users: the C sysctl may need to be increased if +handling thousands of inboxes (with L) to avoid +out-of-memory errors from git. + +Other OSes may have similar tuning knobs (patches appreciated). + +=head2 Scalability to many inboxes + +L allows any number of public-inboxes +to share the same Xapian indices. + +git 2.33+ startup time is orders-of-magnitude faster and uses +less memory when dealing with thousands of alternates required +for thousands of inboxes with L. + +Frequent packing (via L) both improves performance +and reduces the need to increase C. + =head1 CONTACT Feedback encouraged via plain-text mail to L @@ -130,10 +183,10 @@ Information for *BSDs and non-traditional filesystems especially welcome. Our archives are hosted at L, -L, and other places +L, and other places =head1 COPYRIGHT -Copyright 2020 all contributors L +Copyright all contributors L License: AGPL-3.0+ L