Full restore over SFTP: read packs ahead on four connections at once

A restore read one pack at a time over one SFTP connection, which over a long
line is bound by the round trip and by what one TCP connection gets: about
2-3 MiB/s from a Hetzner Storage Box to Canada, with PHP mostly waiting.

Repository::restore() now first lists the ranges of packs the restore needs,
in order (ReadAheadPlan), and the location reads them ahead into unlinked
files in a spool folder in the target (IReadsAhead). Over SFTP,
SftpRangeFetcher does that on four connections of its own, 16 requests in
flight each, taking the answers of whichever connection has one. A range that
fails, a damaged copy, or a blob that moved since the plan is read the usual
way, which reports what is wrong. Other locations restore as before.

On a test line with 120 ms RTT and 0.2% loss, a 1.6 GiB restore went from
5.1 to 17.8 MiB/s, all files identical.
This commit is contained in:
2026-10-19 01:40:49 +02:00
parent b789d17636
commit ccbbf01c54
12 changed files with 1228 additions and 5 deletions
+40
View File
@@ -0,0 +1,40 @@
<?php
declare(strict_types=1);
// SPDX-FileCopyrightText: 2026 Dillard Blom
// SPDX-License-Identifier: AGPL-3.0-or-later
namespace OCA\NgBackup\Backend;
/**
* One range that SftpRangeFetcher reads on one connection.
*
* @internal
*/
final class SftpRangeJob {
public string $handle = '';
/** Offset of the next READ to send. */
public int $next;
/** @var list<array{int, int}> offset and length of parts to ask for again (after a short answer) */
public array $redo = [];
/** @var list<array{int, int, int, int}> request id, kind, offset and length, in the order sent */
public array $queue = [];
/** Bytes received. */
public int $got = 0;
public bool $failed = false;
/** @var resource the spool file that holds the range */
public $fh;
/** @param resource $fh */
public function __construct(
public int $i,
$fh,
public int $start,
public int $end,
) {
$this->fh = $fh;
$this->next = $start;
}
}