RepoMirrors
/
mars
mirror of https://github.com/schoebel/mars
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

marsadm: fix external races on resize

This commit is contained in:
Thomas Schoebel-Theuer 2015-03-16 16:30:12 +01:00
parent 8888571b2c
commit 1f2680dd62
2 changed files with 167 additions and 28 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

101
docu/mars-manual.lyx
View File

@ -6045,6 +6045,88 @@ reference "sec:Scripting-HOWTO"
.
\end_layout
\begin_layout Subsection
Online Resizing during Operation
\end_layout
\begin_layout Standard
You should have LVM or some other means of increasing the physical size
of your disk (e.g.
via firmware of some RAID controllers).
The network must be healthy.
Do the following steps:
\end_layout
\begin_layout Enumerate
Increase your local disks (usually
\family typewriter
/dev/vg/mydata
\family default
)
\emph on
everywhere
\emph default
in the whole cluster.
In order to avoid wasting space, increase them
\emph on
uniformly
\emph default
to the same size (when possible).
The
\family typewriter
lvresize
\family default
tool is documented elsewhere.
\end_layout
\begin_layout Enumerate
Check that all MARS switches are on.
If not, say
\family typewriter
marsadm up mydata
\family default
everywhere.
\end_layout
\begin_layout Enumerate
At the primary:
\family typewriter
marsadm resize mydata
\end_layout
\begin_layout Enumerate
If you have intermediate layers such as iSCSI, you may need some
\family typewriter
iscsiadm
\family default
update or other command.
\end_layout
\begin_layout Enumerate
Now you may increase your filesystem.
This is specific for the filesystem type and documented elsewhere.
\end_layout
\begin_layout Standard
\noindent
\begin_inset Graphics
filename images/lightbulb_brightlit_benj_.png
lyxscale 12
scale 7
\end_inset
Hint: the secondaries will start syncing the increased new part of the underlyin
g primary disk.
In many cases, this is not really needed, because the new junk data just
does not care.
If you are sure and if you know what you are doing, you may use
\family typewriter
marsadm fake-sync mydata
\family default
to abort such unnecessary traffic.
\end_layout
\begin_layout Section
The State of MARS
\begin_inset CommandInset label
@ -23328,28 +23410,37 @@ status open
\begin_layout Plain Layout
\size scriptsize
Precondition: all disks in the cluster participating in
Precondition: The local host must be primary.
All disks in the cluster participating in
\family typewriter
$res
\family default
must be physically larger than the logical resource size (e.g.
by use of
must be physically larger than the logical resource size (e.g, by use of
\family typewriter
lvm
\family default
; can be checked by macros
\family typewriter
%disk-size{}
\family default
and
\family typewriter
%resource-size{}
\family default
).
When the optional
\family typewriter
$size
\family default
argument is present, it must be smaller than the minimum of all physical
sizes, but larger than the current logical size.
sizes, but larger than the current logical size of the resource.
\end_layout
\begin_layout Plain Layout
\size scriptsize
Postcondition: at the (future) primary (if any), the logical size of
Postcondition: the logical size of
\family typewriter
/dev/mars/$dev_name
\family default

94
userspace/marsadm
View File

@ -422,6 +422,13 @@ sub check_res {
return $res;
}
sub _get_mars_size {
my ($cmd, $res) = @_;
my $dev_name = get_link("$mars/resource-$res/device-$host");
my $info = "/sys/devices/virtual/block/mars!$dev_name/size";
return `cat $info` * 512;
}
sub check_sizes {
my ($res, $peer) = @_;
my $logical_size = get_link("$mars/resource-$res/size");
@ -2192,8 +2199,13 @@ sub invalidate_res_phase3 {
}
}
sub resize_res {
my %resize_device_size;
my %resize_old_size;
my %resize_new_size;
sub resize_phase0 {
my ($cmd, $res, $size_arg) = @_;
ldie "mars kernel module is not loaded. This is needed for communication with some other hosts!\n" if !is_module_loaded();
my $new_size = 0;
if ($size_arg) {
$new_size = get_size($size_arg);
@ -2201,39 +2213,70 @@ sub resize_res {
lprint "new size: $new_size bytes\n";
}
check_primary(@_);
my $my_size = get_link("$mars/resource-$res/actsize-$host");
my @actsizes = glob("$mars/resource-$res/actsize-*");
ldie "resource $res has no actsize-* symlinks\n" unless @actsizes;
my $lnk = "$mars/resource-$res/size";
my $old_size = get_link($lnk);
my $min_size = 0;
my $possible_size = 0;
foreach my $actsize (@actsizes) {
my $this_size = get_link($actsize);
if (!$min_size || $this_size < $min_size) {
$min_size = $this_size;
if (!$possible_size || $this_size < $possible_size) {
$possible_size = $this_size;
}
}
lprint "old_size=$old_size\n";
lprint "min_size=$min_size\n";
$new_size = $min_size if !$new_size;
lprint "possible_size=$possible_size\n";
$new_size = $possible_size if !$new_size;
lprint "new_size=$new_size\n";
ldie "new size $new_size is higher than the minimum size of all volumes $min_size" if $new_size > $min_size; # no override with --force possible
# for now, disallow decreasing until some bugs are fixed
ldie "new size $new_size is higher than the possible size (minimum of all volumes) $possible_size" if $new_size > $possible_size; # no override with --force possible
# disallow decreasing
ldie "only increases of the size are possible!\n" if $new_size < $old_size;
ldie "only increases of the size are possible without --force\n" if $new_size <= $old_size && !$force;
foreach my $switch (glob("$mars/resource-$res/todo-*/sync")) {
my $this_switch = get_link($switch);
ldie "sync on '$switch' is switched on -- use marsadm pause-sync to stop\n" unless !$this_switch;
}
my @syncsizes = glob("$mars/resource-$res/syncstatus-$host");
foreach my $syncsize (@syncsizes) {
my $this_size = get_link($syncsize);
ldie "sync on $syncsize has not yet finished: $this_size != $old_size (DANGEROUS FIX: if you know what you are doing, marsadm fake-sync can 'fix' it -- but this may need a full-sync afterwards)\n" unless $this_size == $old_size;
}
foreach my $syncsize (@syncsizes) {
my $this_size = get_link($syncsize);
set_link($new_size, $syncsize);
}
my $waste = $my_size - $new_size;
lwarn "You are wasting $waste bytes locally\n" if $my_size > $new_size;
# remember values
$resize_device_size{$res} = _get_mars_size(@_);
$resize_old_size{$res} = $old_size;
lwarn "internal mismatch between actual device size and resource size: $resize_device_size{$res} != $resize_old_size{$res}\n" unless $resize_device_size{$res} == $resize_old_size{$res};
$resize_new_size{$res} = $new_size;
return 0;
}
sub resize_phase1 {
my ($cmd, $res) = @_;
my $old_size = $resize_old_size{$res} or ldie "bad internal size value\n";
my $new_size = $resize_new_size{$res} or ldie "bad internal size value\n";
# for safety, check again
ldie "only increases of the size are possible!\n" if $new_size < $old_size;
check_primary(@_);
# Mark the primary data / its size as authoritative
my $act_lnk = "$mars/resource-$res/syncstatus-$host";
set_link($new_size, $act_lnk);
finish_links(); # Chance for errors to pop up
# Now set the new resource size
my $lnk = "$mars/resource-$res/size";
set_link($new_size, $lnk);
finish_links();
}
sub resize_phase2 {
my ($cmd, $res) = @_;
my $old_size = $resize_old_size{$res} or ldie "bad internal size value\n";
my $new_size = $resize_new_size{$res} or ldie "bad internal size value\n";
for (;;) {
my $new_device_size = _get_mars_size(@_);
if ($new_device_size == $resize_new_size{$res}) {
lprint "Device size is now $new_device_size.\n";
last;
}
lprint "Device size $new_device_size has not yet reached the new size $resize_new_size{$res}.\n";
if ($new_device_size != $resize_device_size{$res}) {
lwarn "The size has changed, but did not reach the correct value.";
lwarn "Assuming some rounding problems (which may occur at some device types)\n";
last;
}
sleep_timeout();
}
}
sub role_cmd {
@ -4561,7 +4604,12 @@ my %cmd_table =
"When successful, /dev/mars/\$res at the primary will be increased",
"in size. In addition, all secondaries will start an incremental",
"fast full-sync to get the enlarged parts from the primary.",
\&resize_res,
\&resize_phase0,
"check preconditions",
\&resize_phase1,
"set new size",
\&resize_phase2,
"wait for change",
],
"check-resize" => \&ignore_cmd,
"create-md" => \&senseless_cmd,