- 08 Nov, 2012 40 commits
-
-
Philipp Reisner authored
That is necessary since the whole transfer log is per connection(tconn) and not per device(mdev). This bug caused list corruption on the worker list. When a barrier is queued for sending in the context of one device, another device did not see the CREATE_BARRIER bit, and queued the same object again -> list corruption. Signed-off-by:
Philipp Reisner <philipp.reisner@linbit.com> Signed-off-by:
Lars Ellenberg <lars.ellenberg@linbit.com>
-
Philipp Reisner authored
Signed-off-by:
-
Philipp Reisner authored
* drbd-8.3: drbd: O_SYNC gives EIO on ramdisks for some kernels (eg. RHEL6). drbd: send intermediate state change results to the peer Signed-off-by:
-
Philipp Reisner authored
* drbd-8.3: drbd: fix spurious meta data IO "error" drbd: Fixed a race condition between detach and start of resync drbd: fix harmless race to not trigger an ASSERT drbd: Derive sync-UUIDs only from the bitmap-uuid if it is non-zero drbd: Fixed current UUID generation (regression introduced recently, after 8.3.11) Signed-off-by:
-
Philipp Reisner authored
Since version 4.6.1 gcc warns about variables that get a value assigned, but which are never read later on. Signed-off-by:
-
Philipp Reisner authored
With sync-after dependencies, given "lucky" timing of pause/unpause events, and the end of an empty (0 bits set) resync was sometimes not detected on the SyncTarget, leading to a "stalled" SyncSource state. Fixed this by expecting not only "Inconsistent -> UpToDate" but also "Consistent -> UpToDate" transitions for the peer disk state to end a resync. Signed-off-by:
-
Lars Ellenberg authored
Signed-off-by:
-
Philipp Reisner authored
...not only for the first volume Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Lars Ellenberg authored
If no net-options are configured (all on their default), no DRBD_NLA_NET_CONF will be passed to the kernel. The kernel must not require its presence, there is no required option in there. Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Philipp Reisner authored
* The max of minor_count is 255 * In drbdadm count the number of minors, instead of finding the highest minor number * No longer us the magic in the init script Signed-off-by:
-
Philipp Reisner authored
This was a regression recently introduced with commit 7848ddb752c09b6dfd1ddfabb06b69b08aa8f6b9 "drbd: Correctly handle resources without volumes" Signed-off-by:
-
Lars Ellenberg authored
Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Philipp Reisner authored
If we get into the C_BROKEN_PIPE cstate once, the state engine set the thi->t_state of the receiver thread to restarting. But with the while loop in drbdd_init() a new connection gets established. After the call into drbdd() returns immediately since the thi->t_state is not RUNNING. The restart of drbd_init() then resets thi->t_state to RUNNING. I.e. after entering C_BROKEN_PIPE once, the next successful established connection gets wasted. The two parts of the fix: * Do not cause the thread to restart if we detect the issue with the sockets while we are in C_WF_CONNECTION. * Make sure that all actions that would have set us to C_BROKEN_PIPE happen before the state change to C_WF_REPORT_PARAMS. Signed-off-by: -
Andreas Gruenbacher authored
The last data-integrity-alg fix made data integrity checking work when the algorithm was changed for an established connection, but the common case of configuring the algorithm before connecting was still broken. Fix that. Signed-off-by:
Andreas Gruenbacher <agruen@linbit.com> Signed-off-by:
-
Andreas Gruenbacher authored
There is no need to overly generalize this function; it only makes the code harder to understand. Signed-off-by:
-
Philipp Reisner authored
Since we now apply the AL in user space onto the bitmap, the AL is not active for the requests we want to reply. For that a al_write_transaction() that might be called from worker context became necessary. Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Philipp Reisner authored
In case we can not find out why the request takes too long (happens e.g. when IO got suspended on DRBD level). rearm the timer with a reasonable value. Signed-off-by:
-
Philipp Reisner authored
...when the peer has inconsistent data. In that case we failed to clear the susp_nod flag. When the local disk was attached again Signed-off-by:
-
Philipp Reisner authored
Looks like a remainder from long ago. Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher authored
Refer to the settings by the names which drbdsetup and drbd.conf are using. Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher authored
The minor_count module/kernel parameter serves to scale the size of drbd's internal memory pool, but it is no longer a limit for the number of minors or the minor number. (Minor numbers can be arbitrarily high within the allowed limit of 2^20.) Signed-off-by:
-
Andreas Gruenbacher authored
The maximum minor device number allowed by the kernel is (1<<20 - 1). Reject device numbers higher than that to earlier catch possible errors. Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Philipp Reisner authored
* Enabled the resync controller, with a fill target of 50Kib. That gives reasonable resync speeds without tuning. A much better default than the 250KiB/s fixed. * Enable bitmap compression. It is save to use, and most people have more CPU power than network bandwidth. * ko-count of 7: Abort a connection if the peer fails to process a write request within 42 seconds. * al-extents of 1237: ~5 GiB seems to be a much more sane default these days. Signed-off-by:
-
Lars Ellenberg authored
Currently it is legal (though unusual) to create and connect a resource, before adding in all necessary volumes. We should include the network configuration details, even if we don't have a single volume (yet). Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Philipp Reisner authored
When removing a volume/device we need to switch the connection status of the peer back into WFReportParams. Before this fix it was left in Connected state. That means that the peer device continued to inform us about state changes, etc... But we deleted that minor -> protocol error. Signed-off-by:
-
Lars Ellenberg authored
Detection of unclean shutdown has moved into user space. The kernel code will, whenever it updates the meta data, mark it as "unclean", and will refuse to attach to such unclean meta data. "drbdadm up" now schedules "drbdmeta apply-al", which will apply the activity log to the bitmap, and/or reinitialize it, if necessary, as well as set a "clean" indicator flag. This moves a bit code out of kernel space. As a side effect, it also prevents some 8.3 module from accidentally ignoring the 8.4 style activity log, if someone should downgrade, whether on purpose, or accidentally because he changed kernel versions without providing an 8.4 for the new kernel, and the new kernel comes with in-tree 8.3. Signed-off-by:
-
