The commit is pushed to "branch-rh7-3.10.0-862.20.2.vz7.73.x-ovz" and will
appear at https://src.openvz.org/scm/ovz/vzkernel.git
after rh7-3.10.0-862.20.2.vz7.73.6
------>
commit 7d5df8ad0cb2ee138b42683fcd986c89572f462a
Author: Anton Nefedov <[email protected]>
Date: Fri Nov 16 12:46:28 2018 +0300
ploop: move to standby mode after fsync() error too
Extend bdadef81aba5 'ploop: add a standby mode' so it covers fsync().
https://pmc.acronis.com/browse/VSTOR-17613
Signed-off-by: Anton Nefedov <[email protected]>
Acked-by: Andrey Zaitsev <[email protected]>
Acked-by: Sergey Lysanov <[email protected]>
===========================
Patchset description:
ploop: more cases for standby mode
standby mode motivation:
commit bdadef81aba572fdcfd59e4c5c7b18736c962ebd
Author: Andrei Vagin <[email protected]>
Date: Thu Mar 1 11:07:34 2018 +0300
ploop: add a standby mode
This mode shows that a delta lease was stolen and it is impossible to
handle any requests.
We want to know about this situation from the iscsi target. When HA
decides that the current target is broken, it can initialize another
target with the same delta. In this case, the first target has to
complete
all in-porgress commands and set the ASCQ_04H_ALUA_TG_PT_STANDBY bit in
their status.
In Linux, bio-s are always completed with EIO in error cases, so we need
another way how to determine this state. This patch addes a new block
queue flag QUEUE_FLAG_STANDBY.
Anton Nefedov (3):
ploop: move standby mode setting to separate function
ploop: move to standby mode after fsync() error too
ploop: move to standby after -ENOTCONN too
---
drivers/block/ploop/io_kaio.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/block/ploop/io_kaio.c b/drivers/block/ploop/io_kaio.c
index 03d5a471b310..52d3e542fe19 100644
--- a/drivers/block/ploop/io_kaio.c
+++ b/drivers/block/ploop/io_kaio.c
@@ -524,6 +524,7 @@ static int kaio_fsync_thread(void * data)
"on ploop%d)\n",
err, io->files.inode->i_ino,
io2level(io), plo->index);
+ check_standby_mode(err, plo);
PLOOP_REQ_SET_ERROR(preq, -EIO);
} else if (preq->req_rw & REQ_FLUSH) {
BUG_ON(!preq->req_size);
_______________________________________________
Devel mailing list
[email protected]
https://lists.openvz.org/mailman/listinfo/devel
