update_recovery_start() assumed that the out-of-sync disk would always be
marked as IMSM_ORD_REBUILD in the disk_ord_tbl, but the segmentation
fault reported by Andy proves otherwise. This might also be explained by
an interrupted rebuild and the disk has not yet been marked missing.
https://bugzilla.redhat.com/show_bug.cgi?id=592030
Reported-by: Andy Lutomirski <luto@mit.edu>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
This reverts commit fdb482f99b.
Now that containers can report state for ->container_enough we can
automatically determine when the array can be started, and no longer
need the --no-degraded hammer.
Conflicts:
Incremental.c
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
This is needed for imsm where:
1/ we want to report raid_disks as zero to allow mdadm -As to
incorporate all spares
2/ we can't determine stale disks by looking at the event counts.
3/ we can't see per-subarray expectations with the info returned from
the container level ->getinfo_super()
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Normally Monitor doesn't see faulty devices in active slots - they get
moved away too quickly.
But if it does, it reports the "faulty device disappeared" event (when
it finally does get moved away) as SpareActive due to insufficient
checking.
So add a better check.
Reported-by: Pierre Vignéras <pierre@vigneras.name>
A recent change move the sysfs_read call away from the check that it
succeeded. This patch moves the check back next to the sysfs_read
call.
Signed-off-by: NeilBrown <neilb@suse.de>
devnum2devname() returns pointer to memory allocated with strdup.
It must be released to prevent memory leak.
Signed-off-by: Przemyslaw Czarnowski <przemyslaw.hawrylewicz.czarnowski@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
The kernel updates and notifies md/sync_completed when it is time to
take a checkpoint. When this occurs (at 1/16 array size intervals)
write 'idle' to md/sync_action to have the current recovery position
updated in recovery_start and resync_start.
Requires the metadata handler to reset ->last_checkpoint when it has
determined that recovery has ended.
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
When activating a spare we neglect to open recovery_start and as such do
not see checkpoint events. Move disk initialization to common routine
to mitigate recurrence.
Reported-by: Adam Kwolek <adam.kwolek@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Chunks aren't particularly big, but when you could them in bytes
and multiply them together (as we do for calculating the backup
size for 'grow') they can overflow a 32bit int.
So group the division by 512 more closely with the
chunk size so were would need 30Meg chunks to come close to
overflowing 32bits.
Signed-off-by: NeilBrown <neilb@suse.de>
These metadata are not expected on partitions, and they have
no way of differentiation whether which is correct if they
are found both on the device and on the last partition.
So if the device is a partition, refuse to read the metadata.
Signed-off-by: NeilBrown <neilb@suse.de>
mdadm prevents creation when device names are duplicated on the command
line, but leaves the partially created array intact. Detect this case
in the error code from add_to_super() and cleanup the partially created
array. The imsm handler is updated to report this conflict in
add_to_super_imsm_volume().
Note that since neither mdmon, nor userspace for that matter, ever saw an
active array we only need to perform a subset of the cleanup actions.
So call ioctl(STOP_ARRAY) directly and arrange for Create() to cleanup
the map file rather than calling Manage_runstop().
Reported-by: Krzysztof Wojcik <krzysztof.wojcik@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
I suspect commit c132678b wanted VAR_RUN and ALT_RUN to be suffixed by
'/mdadm' to match the defaults in mdadm.h.
Cc: Luca Berra <bluca@comedia.it>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: NeilBrown <neilb@suse.de>
filesystem is writable).
This particularly keeps udev happy if VAR_RUN is set to /dev/md.
Signed-off-by: Doug Ledford <dledford@redhat.com>
Signed-off-by: NeilBrown <neilb@suse.de>
Default metadata type is now 1.2, and we sometimes
add extra alignment before the data section,
so adjust tests for these changes.
Signed-off-by: NeilBrown <neilb@suse.de>
Matching the functionality already in super0 and super1, when
we first create a container, remove any other recognisable metadata to
ensure it doesn't cause confusion.
Signed-off-by: NeilBrown <neilb@suse.de>
Found during testing:
- cannot check metadata for homehost before loading metadata.
- As 1.x metadata can has a state 'rebuilding' between
'spare' and 'ok', we need to include that in our calculations.
Signed-off-by: NeilBrown <neilb@suse.de>
When checking if the new chunk size fit in the component size
we were confusing sectors and K, and so getting it wrong.
Signed-off-by: NeilBrown <neilb@suse.de>
Code to check partition tables used some needless casts
and was broken, using a u8 when a u32 was wanted.
So create structure describing the tables rather than using offset,
and read into those tables instead.
Signed-off-by: NeilBrown <neilb@suse.de>
This reverts commit 9ef5dbff4a as it is
duplicating the check that is done internal to imsm_read_serial().
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
For 1.1 and 1.2 metadata where data_offset is not zero,
it is important to align the data_offset to underlying
block size.
We don't currently have access to the particular device
in avail_size so just try to force to a 1Meg boundary.
Also default 1.x metadata to 1.2 as documented.
Signed-off-by: NeilBrown <neilb@suse.de>
If /dev is static, a name may remain there after the
device has been detached from the system.
Using 'mdadm' to remove such a device from the array
should still work (even though "mdadm --remove detached"
might be preferred).
So when processing a device for '-r', don't insist on
being able to open the device.
Signed-off-by: NeilBrown <neilb@suse.de>
If mdmon is run without "--takeover", it should still
check both *_RUN directories for a valid 'pid' file
to guard against having two mdmons running at the same time.
Suggested-by: Luca Berra <bluca@comedia.it>
Signed-off-by: NeilBrown <neilb@suse.de>
Check that either VAR_DIR or ALT_DIR is actually writable before
selecting it.
Signed-off-by: Luca Berra <bluca@comedia.it>
Signed-off-by: NeilBrown <neilb@suse.de>
- when we waited for the old mdmon to exit, we didn't look
for the socket in the right place
- when we failed to find a pid file, we returned the wrong
value (code expected <0, but got ==0).
Signed-off-by: Luca Berra <bluca@comedia.it>
Signed-off-by: NeilBrown <neilb@suse.de>
running mdmon --takeover --all resulted in a segfault.
Also allow "--all" to be longer so that e.g. md_d127 can
successfully over-write it in the argument list visible
to 'ps'.
Signed-off-by: Luca Berra <bluca@comedia.it>
Signed-off-by: NeilBrown <neilb@suse.de>
When building mdadm.O2, set _FORTIFY_SOURCE to get more
warnings, and also build mdmon.O2 to find warnings in that
code too.
Then fix the warnings.
Suggested-by: Luca Berra <bluca@comedia.it>
Signed-off-by: NeilBrown <neilb@suse.de>
If we find we cannot add the requested bitmap file when
assembling the array, then make sure to clean up properly
and don't leave a half-configured array.
Signed-off-by: NeilBrown <neilb@suse.de>
Change default - again, sorry.
v1.1 causes problems as the conflicts with a boot sector, whether on
the whole device or in a partition. I guess the rule is they you
aren't allowed to use the first sector in general.
v1.2 is mostly just as good and avoid boot-sector problems so default
to that instead.
Signed-off-by: NeilBrown <neilb@suse.de>
As backup file has a timestamp which is updated quite separately
from the metadata timestamp. They should be largely in-sync but
sometimes are not.
So be more generous in the check, and allow it to be over-ridden
by an environment variable.
Signed-off-by: NeilBrown <neilb@suse.de>