When building container members with -IR, we need to ensure that
devices added to an active array preserve the 'in_sync' status so they
don't needlessly get rebuilt.
So allow sysfs_add_disk to do this (only works in kernels since
2.6.30) and pass the relevant flag down.
Signed-off-by: NeilBrown <neilb@suse.de>
I'm not attaching a patch for this because it's so simple. Long story
short, watching both add and change events in udev rules is bad for md
devices. Specifically, the kernel will generate a change event on
things like array stop, and on things like fdisk close. In the case
of array stop, it can result in the array being assembled again
immediately. In the case of fdisk close, the situation is worse.
Let's say you stop all the md devices on some block device in order to
repartition. You run fdisk, change the partition table, then issue a
write of the table. The write of the table triggers the change event
*before* the kernel updates the partition table in memory for the
block device, causing udev to rerun the incremental rules on the old
partition table and restart all the arrays you just stopped with the
old partition table layout, at which point the kernel is unable to
reread the partition table. So, once you've enable incremental
assembly, it becomes apparent that what we really want is to only
start devices on add, not on add|change.
--
Doug Ledford <dledford@redhat.com>
Simple patch to silence some compile warnings that only show up on
64bit arches.
Signed-off-by: Doug Ledford <dledford@redhat.com>
Signed-off-by: NeilBrown <neilb@suse.de>
The 'work_disks' number should be the number that is expected, not the
number found so far. This is needed for Incremental assembly to
start the array at the right time.
Signed-off-by: NeilBrown <neilb@suse.de>
The variable 'i' was being used as a loop variable, and also
for something else inside the loop. So make the larger loop have a
more meaningful name.
Signed-off-by: NeilBrown <neilb@suse.de>
wait not only for the name to appear, but for it to refer to the
correct device.
Sometimes old symlinks left lying around can be confusing.
Signed-off-by: NeilBrown <neilb@suse.de>
When reporting "--detail --scan", use names like /dev/md/foo where
available rather than /dev/md/127
This is particularly needed for containers where the member arrays
will report "container=/dev/md/foo" and we want the container to have
the same name.
Signed-off-by: NeilBrown <neilb@suse.de>
There are probably other places where rounding size to
chunksize is needed, or useful, but this is a good start.
Signed-off-by: NeilBrown <neilb@suse.de>
In some cases we should only print an error message if
'devname' is defined. In fact we were only returning
the error at all in that case!!
Signed-off-by: NeilBrown <neilb@suse.de>
If an array reshape completed within 1 second, then --grow will not
notice that it has finished and will keep waiting for the critical
section to pass.
So be more cautious in the test.
Signed-off-by: NeilBrown <neilb@suse.de>
If mdmon sees a device added to a container, it should assume it is
a new spare. It could be a part of the array that just hadn't been
assembled yet. So check first.
Signed-off-by: NeilBrown <neilb@suse.de>
If we haven't got hold of all the devices yet, we need to be
ready to skip over some while gathering content information.
Signed-off-by: NeilBrown <neilb@suse.de>
If we are assembling an array in a container and it isn't complete
enough to start yet, then
- don't start mdmon
- don't say the array is started
- don't wait for the device to appear in /dev
Signed-off-by: NeilBrown <neilb@suse.de>
If we need to add digits to a name to make it unique, but don't have
to add '_', we need to avoid adding a digit immediately after a digit.
So if the last character of the name is a digit, add the '_' anyway.
Signed-off-by: NeilBrown <neilb@suse.de>
1/ if homehost matches, then we need to set trustworthy to 'LOCAL'
2/ if we decide to set trustworthy to 'METADATA' because we have to
use the metadata version name, do that *after* we have checked if
we are going to assemble within a container, as inside the
container there could be different sources of names to use.
Signed-off-by: NeilBrown <neilb@suse.de>
DDF raid6 layouts are subtly different from the standard 'md' layouts.
From 2.6.30 the kernel knows about these.
Teach mdadm about them, and also allow 'ddf' to set an appropriate default.
Signed-off-by: NeilBrown <neilb@suse.de>
The information about how slots and roles in the array lined up
turned out to be confusing.
So simplify it and one provide the interesting information.
Signed-off-by: NeilBrown <neilb@suse.de>
If the sector size is > 512, we need to be more careful about
alignment.
The largest known sector size is 4096 and (fortunately) both the
superblock and (in many cases) the bitmap are 4096-byte aligned.
So there should be no data-overlap problems.
The exception is when the bitmap is squeezed into the 3K after the
superblock. This arrangement cannot currently be supported on
4K sector-size devices.
Signed-off-by: NeilBrown <neilb@suse.de>
.. because some devices (dasd) have 4096 byte sector size.
As the superblock is 4096 bytes and the bitmap is in a
60K region, this is safe from any possible corruption.
Signed-off-by: NeilBrown <neilb@suse.de>
There was a few kernel releases where the kernel would shrink max_dev
to be just enough to hold the current number of devices.
More recent kernels never shrink it.
However to be as compatible as possible, if we notice that
max_dev is too small to successfully add a device, increase it.
Signed-off-by: NeilBrown <neilb@suse.de>
Currently Incremental_container is being called after adding each disk.
In the imsm case where spares are not tracked in the raid_disks field we
can use --no-degraded to block premature assembly.
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Just like the Assemble case, default to the text_version of the
container if another name is not specified.
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
mdmon may miss events because it re-reads state after read_and_act. The
additional read is used to determine dirty status before allowing a
sigterm to proceed. Since read_and_act is in the best position to
determine 'dirty' status and its return value is not used, modify it to
return true if the array is dirty.
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
In support of auto-layout:
1/ collect and merge all extents to find the largest common-start free region
2/ verify that we meet the "all volumes must use the same set of disks"
2/ mark the disks to be added in add_to_super_imsm_volume
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
'subdevs' is read from the container in the auto-layout case so reset
subdevs dependent default values. 'insert_point' without this
change is always 2 blocking creation of arrays with > 2 raid disks.
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
NeilBrown
Doug Ledford
Dan Williams