| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
This is to cope with timeouts when recoveries and transactions collide.
Maybe 100 is too hight, but 10 or even 20 have been too low in a
very busy environment.
Michael
|
|
so that it is correctly handled by recoveries.
Also set the dmaster explicitly.
Michael
|
|
|
|
|
|
right. The previous bugs were due to the fact that get_nt_acl_internal()
could return an NTSTATUS error if there was no stored ACL blob, but
otherwise would return the underlying ACL from the filysystem. Fix
this so it always returns a valid acl if it can, and if it does not
its an error to be reported back to the client. This then changes
the inherit acl code. Previously we were trying to match Windows
by setting a minimal ACL on a new file that didn't inherit anything
from a parent directory. This is silly - the returned ACL wouldn't
match the underlying UNIX permissions. The current code will correctly
inherit from a parent if a parent has any inheritable ACE entries
that apply to the new object, but will return a mapping from the
underlying UNIX permissions if the parent has no inheritable entries.
This makes much more sense for new files/directories.
Jeremy.
|
|
posix_fallocate is more efficient than manual zero'ing the file. When
preallocation in kernel space is supported it's extremely fast. Support for
preallocation at fs layer via posix_fallocate and fallocate at kernel site
can be found in Linux kernel 2.6.23/glibc 2.10 with ext4, XFS and OCFS2. Other
systems that I know of which support fast preallocation in kernel space are
AIX 6.1 with JFS2 and recent Solaris versions with ZFS maybe UFS2, too.
People who have a system with preallocation in kernel space might want to set
"strict allocate = yes". This reduces file fragentation and it's also safer for
setups with quota being turned on.
As of today most systems still don't have preallocation in kernel space, and
that's why "strict allocate = no" will stay the default for now.
|
|
|
|
|
|
|
|
look at the mtime
|
|
|
|
|
|
|
|
|
|
Step 0 to restore it as a per-share paramter
|
|
samba.
Guenther
|
|
Catched by smbconftort test on the buildfarm.
Guenther
|
|
Guenther
|
|
Guenther
|
|
This reverts commit e23d8a3d1f558a7e98ef2afd71e1d15c5b3a71bc.
|
|
explicit"
This reverts commit 1e22899d268ae5a089f941a204413c07ee64fc78.
|
|
This reverts commit f7b4151a64d8c6851e62255a7139fd00a5fc63a3.
|
|
This reverts commit c85a4c9ba4a7de65a7850f6f9708df66bd24deea.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Michael
|
|
Michael
|
|
|
|
To reduce code duplication.
Michael
|
|
in the "user.DOSATTRIB" EA. From the docs:
In Samba 3.5.0 and above the "user.DOSATTRIB" extended attribute has been extended to store
the create time for a file as well as the DOS attributes. This is done in a backwards compatible
way so files created by Samba 3.5.0 and above can still have the DOS attribute read from this
extended attribute by earlier versions of Samba, but they will not be able to read the create
time stored there. Storing the create time separately from the normal filesystem meta-data
allows Samba to faithfully reproduce NTFS semantics on top of a POSIX filesystem.
Passes make test but will need more testing.
Jeremy.
|
|
Found with "dfree command" set
|
|
|
|
ndr_interface_table"
This reverts commit 5fc9d93408effe75abcd231c45cbc14656692ebe.
|
|
|
|
|
|
When something in the cluster blocks, it can happen that we wait indefinitely
long for ctdb, just adding to the blocking condition. In theory, nothing should
block, but as someone said "In practice the difference between theory and
practice is larger than in theory". This adds a timeout parameter in seconds,
after which we stop waiting for ctdb and panic.
|
|
Jeremy.
|
|
smbd just crashed on me: In a debug message I called a routine preparing a
string that itself used debug_ctx. The outer routine also used it after the
inner routine had returned. It was still referencing the talloc context
that the outer debug_ctx() had given us, which the inner DEBUG had already
freed.
|
|
for the case that another local process has started a transaction
bewteen releasing the transaction_lock record and starting the
transaction.
Michael
|
|
in db_ctdb_transaction_fetch_start() for error conditions when re-fetching
the transaction_lock record inside the transaction
Michael
|
|
node.
In ctdb_transaction_commit(), when the trans2_commit control fails, there
is a race condition in the 1 second sleep between the local transaction_cancel
and the call to ctdb_replay_transaction(): The database is not locked, and
neither is the transaction_lock record. So another client can start and possibly
complete a new transaction in this gap, but only on the same node: The locking
of the transaction_lock record on a different node which involves migration of
the record to the other node has been disabled by introduction of the
transaction_active flag on the db which closes precisely this gap from the start
of the commit until the call to TRANS2_FINISH or TRANS2_ERROR.
But this mechanism does not cover the case where a process on the same node
tries to start a transaction: There is no obstacle to locking the transaction_lock
record because the record does not need to be migrated.
This commit closes this race condition in ctdb_transaction_fetch_start()
by using the new ctdb_ctrl_transaction_active() call to ask the local
ctdb daemon whether it has a transaction running on the database.
If so, the check is repeated until the running transaction is done.
This does introduce an additional call to the local ctdbd when starting
transactions, but it does close the (hopefully) last race condition.
Michael
|
|
CTDB_CONTROL_TRANS2_COMMIT
Michael
|
|
There are two races in concurrent transactions on a single node.
One in starting a transaction and one with replay during commit.
This commit closes the first race by storing the client pid in the
transaction-lock record and comparing the stored pid against its own
pid after releasing the lock and refetching the record inside the
transaction.
Michael
|
|
Michael
|
