wroclaw-git-backups/nix
1
0
Fork 0
mirror of https://github.com/NixOS/nix synced 2025-06-27 00:11:17 +02:00
Code Activity

nix-copy-closure: Fix race condition

Browse source 
There is a long-standing race condition when copying a closure to a
remote machine, particularly affecting build-remote.pl: the client
first asks the remote machine which paths it already has, then copies
over the missing paths. If the garbage collector kicks in on the
remote machine between the first and second step, the already-present
paths may be deleted. The missing paths may then refer to deleted
paths, causing nix-copy-closure to fail. The client now performs both
steps using a single remote Nix call (using ‘nix-store --serve’),
locking all paths in the closure while querying.

I changed the --serve protocol a bit (getting rid of QueryCommand), so
this breaks the SSH substituter from older versions. But it was marked
experimental anyway.

Fixes #141.
This commit is contained in:
Eelco Dolstra 2014-07-10 11:51:22 +02:00
parent 2c3a8f787b
commit 04170d06bf
4 changed files with 137 additions and 61 deletions
Download patch file Download diff file Expand all files Collapse all files

77
perl/lib/Nix/CopyClosure.pm
View file

@ -4,6 +4,15 @@ use strict;
use Nix::Config;
use Nix::Store;
use List::Util qw(sum);
use IPC::Open2;
sub readInt {
my ($from) = @_;
my $resp;
sysread($from, $resp, 8) == 8 or die "did not receive valid reply from remote host\n";
return unpack("L<x4", $resp);
}
sub copyTo {
@ -20,14 +29,76 @@ sub copyTo {
# Ignore exit status because this is just an optimisation.
}
# Start ‘nix-store --serve’ on the remote host.
my ($from, $to);
my $pid = open2($from, $to, "ssh $sshHost @{$sshOpts} nix-store --serve --write");
# Do the handshake.
eval {
my $SERVE_MAGIC_1 = 0x390c9deb; # FIXME
my $clientVersion = 0x200;
syswrite($to, pack("L<x4L<x4", $SERVE_MAGIC_1, $clientVersion)) or die;
die "did not get valid handshake from remote host\n" if readInt($from) != 0x5452eecb;
my $serverVersion = readInt($from);
die "unsupported server version\n" if $serverVersion < 0x200 || $serverVersion >= 0x300;
};
if ($@) {
chomp $@;
warn "$@; falling back to old closure copying method\n";
return oldCopyTo(\@closure, @_);
}
# Send the "query valid paths" command with the "lock" option
# enabled. This prevens a race where the remote host
# garbage-collect paths that are already there.
my $req = pack("L<x4L<x4L<x4", 1, 1, scalar @closure);
for my $s (@closure) {
my $len = length $s;
$req .= pack("L<x4", $len);
$req .= $s;
$req .= "\000" x (8 - $len % 8) if $len % 8;
}
syswrite($to, $req) or die;
# Get back the set of paths that are already valid on the remote host.
my %present;
my $n = readInt($from);
while ($n--) {
my $len = readInt($from);
my $s;
sysread($from, $s, $len) == $len or die;
$present{$s} = 1;
sysread($from, $s, 8 - $len % 8) if $len % 8; # skip padding
}
my @missing = grep { !$present{$_} } @closure;
return if !@missing;
# Send the "import paths" command.
syswrite($to, pack("L<x4", 4)) or die;
exportPaths(fileno($to), $sign, @missing);
readInt($from) == 1 or die;
# Shut down the server process.
close $to;
waitpid $pid, 0;
}
# For backwards compatibility with Nix <= 1.7. Will be removed
# eventually.
sub oldCopyTo {
my ($closure, $sshHost, $sshOpts, $storePaths, $compressor, $decompressor,
$includeOutputs, $dryRun, $sign, $progressViewer, $useSubstitutes) = @_;
# Ask the remote host which paths are invalid. Because of limits
# to the command line length, do this in chunks. Eventually,
# we'll want to use ‘--from-stdin’, but we can't rely on the
# target having this option yet.
my @missing = ();
my @missing;
my $missingSize = 0;
while (scalar(@closure) > 0) {
my @ps = splice(@closure, 0, 1500);
while (scalar(@$closure) > 0) {
my @ps = splice(@$closure, 0, 1500);
open(READ, "set -f; ssh $sshHost @{$sshOpts} nix-store --check-validity --print-invalid @ps|");
while (<READ>) {
chomp;