Perfect, thank you!
For what it's worth, the error did not happen with regular, locally
attached NVME devices. It only occurred with NVMEoF devices, with the
chelsio driver shipped with the kernel. You seemed to discuss this a bit in
https://www.redhat.com/archives/dm-devel/2016-June/msg00430.html
On Wed, Oct 31 2018 at 5:32am -0400,
Drew Hastings wrote:
>Firstly, thanks for the hard work you guys are doing on the dm drivers.
>I'm curious to know if I correctly understand the limitations of the
>multipath target. I'm using kernel 4.19.0-rc5
>When attempting to create a dev
When using kcopyd to run callbacks through dm_kcopyd_do_callback() or
submitting copy jobs with a source size of 0, the jobs are pushed
directly to the complete_jobs list, which could be under processing by
the kcopyd thread. As a result, the kcopyd thread can continue running
completed jobs indefi
kcopyd has no upper limit to the number of jobs one can allocate and
issue. Under certain workloads this can lead to excessive memory usage
and workqueue stalls. For example, when creating multiple dm-snapshot
targets with a 4K chunk size and then writing to the origin through the
page cache. Synci
This patchset fixes two kcopyd and dm-snapshot related issues:
1. If kcopyd is not used properly, the kcopyd job slab cache can grow
without limit causing excessive memory usage, which can lead to
user processes being killed by the OOM killer.
2. The kcopyd thread can hog the CPU wh
We should be able to distinguish the case where a checker
determines a path to be positively down from the case where
the checker fails to obtain necessary information, e.g.
because of a configuration problem (wrong checker).
Use PATH_WILD for the latter case, as it's hardly used now.
Provide a ge
Replace the character array "message" in struct checker with
a "message ID" field.
The generic checker code defines a couple of standard message IDs
and corresponding messages. Checker-specific message IDs start
at CHECKER_FIRST_MSG. Checkers that implement specific message
IDs must provide a tabl
Hi Christophe,
This is v4 of my "checkers overhaul" series. In v3, I missed
one minor suggestion from Ben. Changed wrt v3 are only 03/22
and 12/22; I kept Ben's Reviewed-by: tag in 12/22.
This series starts with a few minor fixes and then attempts
an overhaul of the checker code.
First, there's
Firstly, thanks for the hard work you guys are doing on the dm drivers.
I'm curious to know if I correctly understand the limitations of the
multipath target. I'm using kernel 4.19.0-rc5
When attempting to create a device from two NVMEs connected over nvme_rdma
/ nvmf, I get the following message
A checker shouldn't set the path state to PATH_DOWN if it fails
to obtain information about the path in the first place. Add logic
to the checker to distinguish a failed path from an unsupported path.
Signed-off-by: Martin Wilck
---
libmultipath/checkers/emc_clariion.c | 54 +
* Hashtable has been replaced by rbtree to manage buffers.
Update the comment.
* Fix typo in the comment for dm_bufio_issue_flush
Signed-off-by: Shenghui Wang
---
drivers/md/dm-bufio.c | 12 ++--
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/drivers/md/dm-bufio.c b/driv
11 matches
Mail list logo
