FAQ Search Today's Posts Mark Forums Read
» Video Reviews

» Linux Archive

Linux-archive is a website aiming to archive linux email lists and to make them easily accessible for linux users/developers.


» Sponsor

» Partners

» Sponsor


 
 
LinkBack Thread Tools
 
Old 11-10-2011, 10:29 AM
 
Default Multipath.conf

Hi all.

I have some comments regarding multipath.conf.
1) Can someone point me to the latest version of the documentation?
The following examples seem to show that the multipath(d) are not in sync
with the man.
a) polling_interval
The man page and the comments in the multipath.conf say that the default
value is 5, but when I execute "multipath -t" when I didn't specified the
value in multipath.conf it says the polling_interval is 0.
I don't even know what 0 means.
Is "multipath -t" giving incorrect values? Is my reasonning wrong? Other?

b) path_grouping_policy
man says that the default is multibus.
Red Hat Documentation (
http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/6/html/DM_Multipath/config_file_defaults.html
) says the default is Failover.

c) prio
man says default is none.
Red Hat Documentation (
http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/6/html/DM_Multipath/config_file_defaults.html
) says the default is const.

d) Why "prio emc" gives "sda: emc prio: query command indicates error",
when I execute multipath -v2 -d with multipathd not running. And sda is
even in the blacklist.

e) If I have EMC Clariion Disk Array, should I use: "path_checker
emc_clariion"?

f) What is the correct getuid_callout:
- /sbin/scsi_id -g -u -s /block/%n (As old EMC document says)
- /sbin/scsi_id -g -u -s (As the man says)
- "/lib/dev/scsi_id --whitelisted --device=/dev/%n" (As
http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/6/html/DM_Multipath/config_file_defaults.html
says.)

g) There maybe more differences. I didn't check all.

2) Maybe this is not a question to this forum, but what are the correct
defaults section in multipath.conf for a host (SLES 10) in a cluster (so
quick failover), dual HBA to dual switch, with "
# Emulex lpfc options
options lpfc lpfc_nodev_tmo=3 lpfc_devloss_tmo=3
"
in /etc/modprobe.conf.

What do you think of:
defaults {
polling_interval 5
# max_polling_interval 4
udev_dir /dev
verbosity 2
selector "round-robin 0"
path_grouping_policy multibus
getuid_callout /sbin/scsi_id -g -u -s /block/%n
prio const
# prio_callout mpath_prio_emc /dev/%n
features 1 queue_if_no_path
path_checker directio
failback immediate
# Previous one for a quick failover
rr_min_io 1000
rr_weight priorities
no_path_retry fail
# Previous one for a quick failover
# pg_timeout none
user_friendly_names yes
max_fds 8192
bindings_file "/etc/multipath_bindings"
# Previous one for future compatibility and same partition has /root
dev_loss_tmo 2
# Previous one for a quick failover
fast_io_fail_tmo 1
# Previous one for a quick failover
async_timeout 30
}
Because of the first problem I prefer to define the parameters even if they
are equal to the defaults values instead of not defining the parameters.

Thanks in advance, TA.

--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel
 
Old 11-12-2011, 11:18 AM
Christophe Varoqui
 
Default Multipath.conf

> 1) Can someone point me to the latest version of the documentation?
> The following examples seem to show that the multipath(d) are not in sync
> with the man.

The uptodate information is the one shipped with the sources. It may
have rotten parts ... but feel free to point them out for us to fix.

> a) polling_interval
> The man page and the comments in the multipath.conf say that the default
> value is 5, but when I execute "multipath -t" when I didn't specified the
> value in multipath.conf it says the polling_interval is 0.
> I don't even know what 0 means.
> Is "multipath -t" giving incorrect values? Is my reasonning wrong? Other?
>
Fixed as of today by commit 56753df50b08d67fb711e215a556d005f82ddf7b

> b) path_grouping_policy
> man says that the default is multibus.
> Red Hat Documentation (
> http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/6/html/DM_Multipath/config_file_defaults.html
> ) says the default is Failover.
>
I have no such default in either multipath.8 multipathd.8 and
multipath.conf.defaults

multipath.conf.annotated correctly points failover as the default policy

> c) prio
> man says default is none.
> Red Hat Documentation (
> http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/6/html/DM_Multipath/config_file_defaults.html
> ) says the default is const.
>
I have no such default in either multipath.8 multipathd.8 and
multipath.conf.defaults has the correct const default.

Fixed in multipath.conf.annotated though

> d) Why "prio emc" gives "sda: emc prio: query command indicates error",
> when I execute multipath -v2 -d with multipathd not running. And sda is
> even in the blacklist.
>
> e) If I have EMC Clariion Disk Array, should I use: "path_checker
> emc_clariion"?
>
Depends on the Clariion settings. Either that or alua

> f) What is the correct getuid_callout:
> - /sbin/scsi_id -g -u -s /block/%n (As old EMC document says)
> - /sbin/scsi_id -g -u -s (As the man says)
> - "/lib/dev/scsi_id --whitelisted --device=/dev/%n" (As
> http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/6/html/DM_Multipath/config_file_defaults.html
> says.)
>
The correct one is the one that works on your system. Different scsi_id
version running on different kernel version have different behaviour.
The internal default makes it work with latest kernel/scsi_id

> g) There maybe more differences. I didn't check all.
>
Please do, if you have the means. Input is valuable.


--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel
 

Thread Tools




All times are GMT. The time now is 09:52 PM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.
Copyright 2007 - 2008, www.linux-archive.org