For non-existent maps (ENXIO from ioctl()), dm_task_run and
dm_task_get_info return success. We need to check info.exists.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
kpartx/devmapper.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/kpartx/devmapper.c b/
This function is not used any more.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
kpartx/lopart.c | 31 ---
kpartx/lopart.h | 1 -
2 files changed, 32 deletions(-)
diff --git a/kpartx/lopart.c b/kpartx/lopart.c
index 2eb3f631..44f0c277 100644
--- a/
arted only after
udev settle and thus the modules are loaded too late for udev
rule processing, causing "multipath" invocations from udev rules to fail
and paths to be wrongly classified as non-multipath.
Fixes: d7188fcd "multipathd: start daemon after udev trigger"
Si
ince
> librados isn't supported on all architectures yet). If people have
> objections, I can continue to use the existing build option.
Well, we're getting close to having to write configure.ac,
or something of the kind :-)
But it's fine with me. ACK from me for this series.
Martin
--
On Wed, 2017-09-13 at 15:53 -0500, Benjamin Marzinski wrote:
> On Sun, Sep 03, 2017 at 12:38:44AM +0200, Martin Wilck wrote:
> > When we process udev rules, it's crucial to know whether I/O on a
> > given
> > device will succeed. Unfortunately DM_NR_VALID_PATHS is not
>
On Wed, 2017-09-13 at 15:33 -0500, Benjamin Marzinski wrote:
> On Sun, Sep 03, 2017 at 12:38:38AM +0200, Martin Wilck wrote:
> > Some vendor kernels (e.g. SUSE) have supported loading multipath
> > maps without valid paths for a long time. Without that feature,
> > problems
Hi Ben,
On Wed, 2017-09-13 at 16:19 -0500, Benjamin Marzinski wrote:
> On Sun, Sep 03, 2017 at 12:38:50AM +0200, Martin Wilck wrote:
> > If DM_ACTIVATION is set by the general dm rules, we need to
> > bring up this device. But if the mpath device is not ready,
> > th
have a test procedure to derive the optimal
settings experimentally, thus guesswork is going to be applied, with
questionable odds for success.
IOW: the whole stuff is basically useless without good default values.
It would be up to you hardware guys to come up with them.
> san_path_err_for
On Fri, 2017-09-15 at 15:40 -0500, Benjamin Marzinski wrote:
> On Thu, Sep 14, 2017 at 03:06:43PM +0200, Martin Wilck wrote:
> > Hi Ben,
> >
> > On Wed, 2017-09-13 at 16:19 -0500, Benjamin Marzinski wrote:
> > > On Sun, Sep 03, 2017 at 12:38:
hen
> devices are
> configured to use the prkeys file, and will set and clear the
> reservation
> keys automatically.
>
> Changes in v2:
> - use struct be64 instead of uint64_t to hold the reservation_key
>(Martin Wilck)
>
> Benjamin Marzinski (5):
> libmu
On Wed, 2017-09-20 at 13:32 +0200, Martin Wilck wrote:
> On Fri, 2017-09-15 at 08:15 +0200, Hannes Reinecke wrote:
> >
> > Signed-off-by: Hannes Reinecke <h...@suse.com>
> > ---
> > libmultipath/checkers.c | 4 ++--
> > libmultipath/print.c| 2 +-
&
On Wed, 2017-09-13 at 16:19 -0500, Benjamin Marzinski wrote:
> On Sun, Sep 03, 2017 at 12:38:49AM +0200, Martin Wilck wrote:
> > The fact alone that a map changes from not ready to ready does
> > not imply that it is activating.
>
> NAK on this one and
> [PATCH 21/31] 11-
On Wed, 2017-09-13 at 16:19 -0500, Benjamin Marzinski wrote:
> On Sun, Sep 03, 2017 at 12:38:49AM +0200, Martin Wilck wrote:
> > The fact alone that a map changes from not ready to ready does
> > not imply that it is activating.
>
> NAK on this one and
> [PATCH 21/31] 11-
, so we can rely on the latter to create
the symlinks.
The reason I put this last in the series is that it will possibly
require changes in other packages, notably dracut, in order to make
sure partitions mappings are cleanly set up during boot.
Signed-off-by: Martin Wilck <mwi...@suse.
this behavior are provided via kernel parameter
"dont_del_part_nodes", or a custom udev rules file setting the
"DONT_DEL_PART_NODES" environment variable.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
kpartx/Makefile | 2 ++
kpartx/del-part-nodes.rules |
exec'ing multipathd (Ben Marzinski)
- pass more parameters from main program, preparing for other commands
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
multipath/main.c | 62
1 file changed, 62 insertions(+)
diff --git a
gt; %s\n",
> > monitored_count, is_uevent_busy()? "True"
> :
> > "False");
>
> Signed-off-by: Hannes Reinecke <h...@suse.com>
> ---
> libmultipath/checkers.c | 4 ++--
> libmultipath/print.c| 2 +-
> 2 files
On Fri, 2017-09-15 at 08:14 +0200, Hannes Reinecke wrote:
> The udev monitor from uevent_listen() needs to be cleaned up from
> a pthread cleanup hook.
> Valgrind complained here.
>
> Signed-off-by: Hannes Reinecke <h...@suse.com>
Reviewed-by: Martin Wilck <mwi...@suse.co
On Thu, 2017-09-07 at 16:57 -0500, Benjamin Marzinski wrote:
> On Tue, Aug 29, 2017 at 12:05:36AM +0200, Martin Wilck wrote:
> > Some multipath commands are dangerous to run while multipathd is
> > running.
> > For example, "multipath -r" may apply a modified
changed, 43 insertions(+), 36 deletions(-)
>
Reviewed-by: Martin Wilck <mwi...@suse.com>
--
Dr. Martin Wilck <mwi...@suse.com>, Tel. +49 (0)911 74053 2107
SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Graham Norton
HRB 21284 (AG Nürnberg)
--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel
_reskey_from_uint64(r, x) do { (r)->_v = (x); } while (0)
Using be64_to_cpu is all too easily forgotten, which leads to nasty
bugs.
Apart from that, ack.
--
Dr. Martin Wilck <mwi...@suse.com>, Tel. +49 (0)911 74053 2107
SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Graham Norton
HRB
amin Marzinski <bmarz...@redhat.com>
> ---
> libmpathpersist/mpath_persist.c | 9 -
> libmpathpersist/mpath_updatepr.c | 33 +-
> ---
> libmpathpersist/mpathpr.h | 2 +-
> 3 files changed, 22 insertions(+), 22 deletions(-)
>
Reviewed-by
libmpathpersist/mpathpr.h| 1 +
> 3 files changed, 37 insertions(+), 16 deletions(-)
>
Reviewed-by: Martin Wilck <mwi...@suse.com>
--
Dr. Martin Wilck <mwi...@suse.com>, Tel. +49 (0)911 74053 2107
SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Graham Norton
cli_handlers.c | 82 +++
> multipathd/cli_handlers.h | 3 +
> multipathd/main.c | 3 +
> 15 files changed, 388 insertions(+), 20 deletions(-)
> create mode 100644 libmultipath/prkey.c
> create mode 100644 libmultipath/prkey.h
Looks ok to me. I suppo
On Tue, 2017-09-05 at 11:54 +0800, Guan Junxiong wrote:
> Cc: Martin Wilck <mwi...@suse.com>
> Signed-off-by: Junxiong Guan <guanjunxi...@huawei.com>
> ---
> libmultipath/uevent.c | 6 +++---
> 1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/lib
On Tue, 2017-09-05 at 10:59 +0800, Guan Junxiong wrote:
> Hi Martin,
> Thanks for you remarks and compiling check.
> An new patch will be updated in this week.
> Other response is inline.
>
> Cheers
> Guan Junxiong
>
> On 2017/9/5 4:59, Martin Wilck wrote:
> &g
hardware table
>
> libdmmp/Makefile | 1 +
> libmultipath/hwtable.c | 27 ++-
> 2 files changed, 23 insertions(+), 5 deletions(-)
I'd appreciate some information or references where you gathered the
information about and tuning recommendations for the new de
changes
https://www.redhat.com/archives/dm-devel/2017-September/msg00275.html
Martin Wilck
multipath: delegate dangerous commands to multipathd (v2)
https://www.redhat.com/archives/dm-devel/2017-September/msg00182.html
Also, please consider merging the following bug fix
On Sun, 2017-08-20 at 02:47 +0200, Xose Vazquez Perez wrote:
> Reported-by: Martin Wilck <mwi...@suse.com>
> Cc: Christophe Varoqui <christophe.varo...@opensvc.com>
> Cc: device-mapper development <dm-devel@redhat.com>
> Signed-off-by: Xose Vazquez Perez <x
ormal, which means , when path checker
> detects it
> is up, it will reinstate into the usable path.
>
> However, how about we schedule the intermittent IO checking process
> again when
> the path_io_err_recovery_time seconds expires. If the number of IO
> erros is less
> than path_io_err_num_t
are
> diff --git a/multipathd/main.c b/multipathd/main.c
> index 4be2c579..d12e6fae 100644
> --- a/multipathd/main.c
> +++ b/multipathd/main.c
> @@ -84,6 +84,7 @@ int uxsock_timeout;
> #include "cli_handlers.h"
> #include "lock.h"
> #include "wai
fer to use equivalent multipathd client commands instead.
For now, do this only for "multipath -r", but other invocations
may be added in the future. Perhaps some day, all "multipath"
commands will be mapped to multipathd actions.
Signed-off-by: Martin Wilck <mwi...@sus
The case f == NULL is already handled by the code from
b1ecdd46b6ec "segment faulty occured in add_feature()" above.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
libmultipath/structs.c | 27 ---
1 file changed, 12 insertions(+), 15 deletion
y one
strncpy() call to make the code easier to review.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
libmultipath/structs.c | 64 +-
1 file changed, 21 insertions(+), 43 deletions(-)
diff --git a/libmultipath/structs.c b/libmultipat
of traffic on dbus as
systemd forwards these messages on the system bus. This can be
seen with systemd.log_level=debug. Better skip these notifications
that don't provide useful information to the user, who is very
unlikely to catch the daemon in "running" state anyway.
Signed-off-by: Martin
On Tue, 2017-08-29 at 08:37 +0200, Hannes Reinecke wrote:
> On 08/29/2017 12:05 AM, Martin Wilck wrote:
> > The case f == NULL is already handled by the code from
> > b1ecdd46b6ec "segment faulty occured in add_feature()" above.
> >
> >
> At one poi
-by: Martin Wilck <mwi...@suse.com>
---
kpartx/kpartx.rules | 10 +-
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/kpartx/kpartx.rules b/kpartx/kpartx.rules
index 64d550de..ea34e449 100644
--- a/kpartx/kpartx.rules
+++ b/kpartx/kpartx.rules
@@ -5,12 +5,12 @@
#
events
are another.
The LVM rules rely on DM_ACTIVATION, but other rules such as
kpartx can't do that, because they'd miss e.g. partition table
changes.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
kpartx/kpartx.rules | 3 +++
multipath/11-dm-mpath.rules | 8 +++-
2 files c
This function is simply factored out from dm_get_maps.
No functional difference.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
libmultipath/devmapper.c | 43 +++
libmultipath/devmapper.h | 1 +
2 files changed, 28 insertions(+), 16 deletions(-)
This code is run before 13-dm-disk.rules which calls "blkid",
therefore the code can be simplified because the condition checked
always holds.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
multipath/11-dm-mpath.rules | 14 +++---
1 file changed, 7 insertions(+), 7 de
We have code for importing blkid variables from db but this code
is missing the _ENC variants needed for creating symlinks, and
ID_FS_LABEL.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
multipath/11-dm-mpath.rules | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff
for a
> +path will last for \fIpath_io_err_sample_time\fR. If the rate of IO
> error on
> +a particular path is greater than the
> \fIpath_io_err_rate_threshold\fR, then
> +the path will not reinstate for \fIpath_io_err_rate_threshold\fR
> seconds unless
> +there is only one activ
> > path_io_err_num_threshold, then the path will not reinstate for
> >
> > It would be more user-friendly to allow the user to specify the
> > error
> > rate threshold as a percentage
>
> Error rate threshold as a percentage is not enough to distinguish
> sma
Junxiong Guan <guanjunxi...@huawei.com>
> > > ---
> > > libmultipath/Makefile | 3 +-
> > > libmultipath/config.h | 1 +
> > > libmultipath/udevrule.c | 191
> > >
> > > libmultipath/udevr
Added a test case for "kpartx: fix part deletion without partition
table".
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
kpartx/test-kpartx | 14 ++
1 file changed, 14 insertions(+)
diff --git a/kpartx/test-kpartx b/kpartx/test-kpartx
index 35b0f064..60b3eb23 100
The LVM rules do this since 2.02.173 (c48149cf80 "udev: also
create /dev/disk/by-part{label,uuid} and gpt-auto-root symlinks").
We have to do it here for partitions on unaccessible mpath maps.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
kpartx/kpartx.rules | 6 ++
1
, so we can rely on the latter to create
the symlinks.
The reason I put this last in the series is that it will possibly
require changes in other packages, notably dracut, in order to make
sure partitions mappings are cleanly set up during boot.
Signed-off-by: Martin Wilck <mwi...@suse.
kpartx.rules increased link priority unconditionally, but for
explicitly marked low prio devices that shouldn't be done.
Fix that. Also, use "+=" for OPTIONS, as most other rules do.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
kpartx/kpartx.rules | 2 +-
1 file changed, 1
Changing dm_get_info and dm_get_uuid for now.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
libmultipath/devmapper.c | 4 ++--
libmultipath/devmapper.h | 4 ++--
2 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/libmultipath/devmapper.c b/libmultipath/devmapper.c
index 7a
Call the factored-out function to obtain the udev device.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
libmultipath/configure.c | 23 +++
1 file changed, 11 insertions(+), 12 deletions(-)
diff --git a/libmultipath/configure.c b/libmultipath/configure.c
index 73
The partition_delimiter setting was effecitvely ignored. Fix it.
Fixes: 95bf339bb9d7 "correctly set partition delimiter on rename"
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
libmultipath/devmapper.c | 10 ++
1 file changed, 6 insertions(+), 4 deletion
If DM_ACTIVATION is set by the general dm rules, we need to
bring up this device. But if the mpath device is not ready,
that would be dangerous; it could hang or produce lots of IO
errors. So remember this state, and try to activate when the
map becomes usable later.
Signed-off-by: Martin Wilck
If we are deleting partition mappings, there's no need to
parse the partition table. We just look for mappings created
by kpartx and destroy them. Without this patch, kpartx fails
to delete partition mappings on devices on which the partition
table has been destroyed.
Signed-off-by: Martin Wilck
This flag can used to avoid any IO on the device itself. Useful
for getting the path state without the risk of hanging or running
into IO errors.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
libmultipath/discovery.c | 8
libmultipath/discovery.h | 2 ++
2 files chang
to do IO).
Moreover, the previous logic for running IMPORT{db} was wrong (condition
used '=="?*"' rather than '!="?*"').
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
kpartx/kpartx.rules | 24 ++--
1 file changed, 18 insertions(+), 6 deletio
this behavior are provided via kernel parameter
"dont_del_part_nodes", or a custom udev rules file setting the
"DONT_DEL_PART_NODES" environment variable.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
kpartx/Makefile | 1 +
kpartx/del-part-nodes.rules |
strip_slash() is copied from kpartx.c, and will be removed there
in a follow-up patch. The others are new helpers.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
kpartx/devmapper.c | 44 ++--
1 file changed, 38 insertions(+), 6 deletions(-)
diff
pretending an
existing partition table although there was none.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
kpartx/dasd.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/kpartx/dasd.c b/kpartx/dasd.c
index f50c1bdb..e418d5ac 100644
--- a/kpartx/dasd.c
+++ b/kpartx/dasd.c
The fact alone that a map changes from not ready to ready does
not imply that it is activating.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
multipath/11-dm-mpath.rules | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/multipath/11-dm-mpath.rules b/multipath
seems to be more universal. So, when the search for the
named partition fails, search for the UUID instead and try to rename the
map.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
kpartx/devmapper.c | 108 -
kpartx/devmapper.h
Add test for the functionality implemented in patch
"kpartx: search partitions by UUID, and rename".
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
kpartx/test-kpartx | 36
1 file changed, 36 insertions(+)
diff --git a/kpartx/test-kpar
We need to distinguish the case where a device came up with
ENV{MPATH_DEVICE_READY}=="0" in the first place from the case
where it changed from "ready" to "not ready".
Otherwise, we may save a wrong state in DM_DISABLE_OTHER_RULES_FLAG_OLD.
Signed-off-by:
if DM_ACTIVATION=1; doing so we
would miss events caused e.g. by partition table editing. It's not
necessary to scan for certain events. This will be handled in another
patch.
Currently this works for multipath only, but code for other targets
could be added if desired.
Signed-off-by: Martin Wilck <
, systemd, and
kpartx quite a bit with failover and 0-paths scenarios, successfully.
The changes are less drastic than the stats below suggest.
A considerable part just moves functionality out of existing code
into separate functions in order to use it elsewhere.
Regards,
Martin
Martin Wilck (31):
libm
... for devices that have been identified as such.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
multipath/multipath.rules | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/multipath/multipath.rules b/multipath/multipath.rules
index 86defc0c..bc1a8529 100644
--- a/mul
... and their partitions. This rule generates irritiating symlinks
like /dev/disk/by-id/raid-${VG}-${LV} for LVM LVs otherwise.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
kpartx/kpartx.rules | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/kpartx/kpartx.rules b/
if a given
map has usable paths. Obviously this command must avoid doing any I/O
on the multipath map itself, thus no checkers are called; only status
from sysfs and dm is collected.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
libmultipath/config.h | 1 +
multipath/
For kernel-generated path events, DM_NR_VALID_PATHS indicates
whether usable paths are available. But this information isn't
reliable as events may be received out of order.
Use multipath -U to determine whether a multipath map can
handle I/O.
Signed-off-by: Martin Wilck <mwi...@suse.
This factors out the functionality of retrieving the
udev_device from different device types from get_refwwid.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
libmultipath/configure.c | 35 +++
libmultipath/configure.h | 1 +
2 files changed, 36 inse
ee whether or not the device is a dm device.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
kpartx/devmapper.c | 44 +++-
kpartx/devmapper.h | 13 +
kpartx/kpartx.c| 25 -
3 files changed, 56 insertions(+), 26 de
On Mon, 2017-10-09 at 14:09 -0500, Benjamin Marzinski wrote:
> On Thu, Oct 05, 2017 at 10:25:50PM +0200, Martin Wilck wrote:
> >
> > Like this, perhaps?
> >
> > Martin Wilck (1):
> > limbultipath: prefer RDAC checker with detect_checker
>
> Oops. I tota
With "detect_checker yes", ALUA is used for all storage devices
that support ALUA. But currently RDAC is still preferred for RDAC
devices
(https://www.redhat.com/archives/dm-devel/2017-September/msg00326.html)
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
libmultipat
). It is set
> > > on by
> > > default just like detect_prio and retain_attached_hw_handler.
> >
> > RDAC devices are going also to be affected with this change.
> > Is it desired?
>
> No. I'll send a patch that adds a check_rdac() call before setting
>
Srivastava <ritika.srivast...@oracle.com>
OK, but IMO it could be done more nicely. I'll post an alternative
patch soon.
Regards,
Martin
--
Dr. Martin Wilck <mwi...@suse.com>, Tel. +49 (0)911 74053 2107
SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Graham Norton
HRB 21284 (
than I do.
Just in case I made a silly mistake.
Martin
--
Dr. Martin Wilck <mwi...@suse.com>, Tel. +49 (0)911 74053 2107
SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Graham Norton
HRB 21284 (AG Nürnberg)
--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mail
n. At openSUSE, we simply set
OPTFLAGS=%{optflags} in the spec file. RPM_OPT_FLAGS aka %{optflags} is
a standard for RPM-based distros, but "OPTFLAGS" is not.
Martin
--
Dr. Martin Wilck <mwi...@suse.com>, Tel. +49 (0)911 74053 2107
SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Graham Norton
HRB 21284 (AG Nürnberg)
--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel
On Wed, 2017-12-13 at 12:08 -0600, Benjamin Marzinski wrote:
> With the fix to the features handling code, the multipath device
> should
> already be reloaded with the correct value for queue_if_no_path, so
> there's no need to go reset it immediately afterwards.
>
> Cc:
with group_by_prio.
Putting Hannes and Ben on CC because I'd like to get their opinion,
too.
Regards
Martin
--
Dr. Martin Wilck <mwi...@suse.com>, Tel. +49 (0)911 74053 2107
SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Graham Norton
HRB 21284 (AG Nürnberg)
--
dm-devel mailing list
dm
Hello Guan,
> On 2017/11/18 8:11, Martin Wilck wrote:
> > The log standard deviation can be calculated much more simply
> > by realizing
> >
> >sum_n (x_i - avg(x))^2 == sum_n x_i^2 - n * avg(x)^2
> >
>
> I derive the equation:
> sum_n {(x_i - av
This function should return 1 on success, as the other sysfs_get_XXX
functions. The callers actually expect this - therefore timeout values
from sysfs are never used.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
libmultipath/discovery.c | 2 +-
1 file changed, 1 insertion(+), 1 de
I don't think anyone can measure latency to 1% accuracy. It's
better to not even pretend to be able to. 10% should be fine
even for the most latency-critical environments.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
libmultipath/prioritizers/path_latency.c | 5 +++--
1 file chan
ne a differentiation, in particular
in combination with group_by_prio, because we'd loose the ability for
load balancing. But this is rather a topic for the man page or a
"best practices" document.
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
libmultip
The log standard deviation can be calculated much more simply
by realizing
sum_n (x_i - avg(x))^2 == sum_n x_i^2 - n * avg(x)^2
Also, use timespecsub rather than the custom timeval_to_usec,
and avoid taking log(0).
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
libmultipath/priori
ease check these fixes.
I also have remaining issues with the "intermittent IO accounting"
patch, will send these later.
Martin Wilck (4):
libmultipath: path latency: fix default base num
libmultipath: path latency: log threshold with p2
libmultipath: path latency: simplify getprio()
This is not a critical error. It just means that the path in
question will have low priority (rightly so, if it has >100s latency).
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
libmultipath/prioritizers/path_latency.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
d
etting
> the MPATH_UDEV_NO_PATHS_FLAG.
>
> Signed-off-by: Benjamin Marzinski <bmarz...@redhat.com>
Reviewed-by: Martin Wilck <mwi...@suse.com>
--
Dr. Martin Wilck <mwi...@suse.com>, Tel. +49 (0)911 74053 2107
SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Graham N
jamin Marzinski <bmarz...@redhat.com>
> ---
Reviewed-by: Martin Wilck <mwi...@suse.com>
--
Dr. Martin Wilck <mwi...@suse.com>, Tel. +49 (0)911 74053 2107
SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Graham Norton
HRB 21284 (AG Nürnberg)
--
dm-devel mailing list
dm-dev
m_map_present() is pointless in __setup_multipath(),
> since the existence of the map was just checked in dm_get_info().
>
> Signed-off-by: Benjamin Marzinski <bmarz...@redhat.com>
> ---
Reviewed-by: Martin Wilck <mwi...@suse.com>
--
Dr. Martin Wilck <mwi...@suse.com>, T
ame thing
> for
> faulty paths.
>
> Signed-off-by: Benjamin Marzinski <bmarz...@redhat.com>
> ---
Reviewed-by: Martin Wilck <mwi...@suse.com>
--
Dr. Martin Wilck <mwi...@suse.com>, Tel. +49 (0)911 74053 2107
SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Gra
ardware handler to the table should trigger a reload in this case.
>
> Signed-off-by: Benjamin Marzinski <bmarz...@redhat.com>
Reviewed-by: Martin Wilck <mwi...@suse.com>
--
Dr. Martin Wilck <mwi...@suse.com>, Tel. +49 (0)911 74053 2107
SUSE Linux GmbH, GF: Felix Imendörffer,
> so
> the output when you created a map was incorrect.
>
> Signed-off-by: Benjamin Marzinski <bmarz...@redhat.com>
I had something similar in mind, but this is actually cleaner.
Reviewed-by: Martin Wilck <mwi...@suse.com>
--
Dr. Martin Wilck <mwi...@suse.com>, Tel
On Thu, 2017-12-07 at 12:48 -0600, Benjamin Marzinski wrote:
> The current del-part-nodes rules try to run partx on the partitions
> themselves, which will ofen fail with an error in the log, because
> the
> partitions will have been deleted before partx can run on them.
>
> Cc:
>
Reviewed-by: Martin Wilck <mwi...@suse.com>
--
Dr. Martin Wilck <mwi...@suse.com>, Tel. +49 (0)911 74053 2107
SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Graham Norton
HRB 21284 (AG Nürnberg)
--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel
ed memory, child() should wait
> until these threads are stopped before calling
> stop_io_err_stat_thread().
>
> Cc: Guan Junxiong <guanjunxi...@huawei.com>
> Signed-off-by: Benjamin Marzinski <bmarz...@redhat.com>
> ---
Reviewed-by: Martin Wilck <mwi...@suse.com>
--
atures,
> - "queue_if_no_pat
> h");
> - }
> - }
AFAICS we could also get rid of the calls to dm_queue_if_no_path() in
reload_map(), right?
Martin
--
Dr. Martin Wilck <mwi...@suse.com>, Tel. +49 (0)911 74053 2107
SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Graham Norton
HRB 21284 (AG Nürnberg)
--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel
which will be removed before
> the
> end of configure(). Now, map_discovery() just gets the necessary
> kernel
> information to make the correct call in select_action().
>
> Signed-off-by: Benjamin Marzinski <bmarz...@redhat.com>
Reviewed-by: Martin Wilck <mwi...@suse.c
accuracy of the
measurement (indicating that 68%, or error_function (sqrt(1/2)), of the
measured values lie in the interval [avg/F, avg*F]). IOW, it tells you
how stable your latencies for a certain path are.
It has nothing to do with the artificial prio "bins" that the latency
prioritizer
On Thu, 2017-12-07 at 12:49 -0600, Benjamin Marzinski wrote:
> setup multipath is only called by the daemon, so there is no point in
> the is_daemon parameter.
>
> Signed-off-by: Benjamin Marzinski <bmarz...@redhat.com>
> ---
Reviewed-by: Martin Wilck <mwi...@suse.com>
>
> Signed-off-by: Benjamin Marzinski <bmarz...@redhat.com>
> ---
Reviewed-by: Martin Wilck <mwi...@suse.com>
--
Dr. Martin Wilck <mwi...@suse.com>, Tel. +49 (0)911 74053 2107
SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Graham Norton
HRB 21284 (AG Nürnberg)
--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel
THS_STRICT, having
the highest numeric value, must be last in the enum.
Fixes: c36f2f42 "libmultipath: change find_multipaths option to multi-value"
Signed-off-by: Martin Wilck <mwi...@suse.com>
---
libmultipath/structs.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff
201 - 300 of 2117 matches
Mail list logo
