All of lore.kernel.org
 help / color / mirror / Atom feed
From: Eric <iggiggitoo@gmail.com>
To: device-mapper development <dm-devel@redhat.com>
Subject: Re: Multipath not using multiple NICs at once
Date: Tue, 25 Mar 2014 09:59:44 -0400	[thread overview]
Message-ID: <53318BD0.7090602@gmail.com> (raw)
In-Reply-To: <532EE739.6090105@bulbous.org>

I tried "service-time 0", but it didn't seem to change. I ran "multipath 
-ll" and it still shows policy="round-robin 0". I restarted the 
multipath service, but no luck.

Here's my current config after some suggested changes from the 
#device-mapper channel on freenode:

defaults {
         path_grouping_policy    multibus
         path_selector           readsector0
         polling_interval        3
         path_selector           "service-time 0"
         failback                immediate
         features                "0"
         no_path_retry           1
         rr_weight               uniform
         rr_min_io_rq            1
#       user_friendly_names     yes
}


On 3/23/2014 9:52 AM, urgrue wrote:
> > path_selector "round-robin 0"
>
> Have you tried changing this to queue-length or service-time?
>
> On 23/3/14 05:08, Eric wrote:
>> Hello,
>>
>> I'm fairly new to multipath and I am having an issue with it not 
>> using all of my NICs. Currently, my node has 4 gigE NICs to my 
>> storage network and the SAN has 8 gigE NICs to the same network and I 
>> am attempting to setup multipath with ISCSI in order to utilize more 
>> than 1 gigabit connection. However, when I use nload to check the 
>> network usage, you can see the traffic hop around the NICs. For 
>> example, data would send for 2-3 seconds on eth1, then stops and 
>> starts on eth2, then stops and starts back up on eth3. All perfectly 
>> distributed, but in this setup, unable to reach beyond the capacity 
>> of a 1 gigabit connection.
>>
>> I have each NIC on a different network (e.g. 10.1.1.0/24 for eth1, 
>> 10.1.2.0/24 for eth2, etc.). Netstat shows that the connections are 
>> being made each to different IPs:
>>
>> tcp        0      0 10.1.3.8:35493          10.1.5.241:3260 ESTABLISHED
>> tcp        0      0 10.1.3.8:53972          10.1.3.241:3260 ESTABLISHED
>> tcp        0      0 10.1.6.8:41090          10.1.4.241:3260 ESTABLISHED
>> tcp        0      0 10.1.1.8:50754          10.1.1.241:3260 ESTABLISHED
>> tcp        0      0 10.1.6.8:49780          10.1.5.241:3260 ESTABLISHED
>> tcp        0      0 10.1.1.8:36938          10.1.6.241:3260 ESTABLISHED
>> tcp        0      0 10.1.6.8:52009          10.1.6.241:3260 ESTABLISHED
>> tcp        0      0 10.1.5.8:51630          10.1.1.241:3260 ESTABLISHED
>> tcp        0      0 10.1.5.8:54481          10.1.4.241:3260 ESTABLISHED
>> tcp        0      0 10.1.1.8:54504          10.1.5.241:3260 ESTABLISHED
>> tcp        0      0 10.1.5.8:58229          10.1.3.241:3260 ESTABLISHED
>> tcp        0      0 10.1.3.8:49031          10.1.1.241:3260 ESTABLISHED
>> tcp        0      0 10.1.5.8:40551          10.1.6.241:3260 ESTABLISHED
>> tcp        0      0 10.1.4.8:45016          10.1.5.241:3260 ESTABLISHED
>> tcp        0      0 10.1.4.8:55665          10.1.4.241:3260 ESTABLISHED
>> tcp        0      0 10.1.6.8:57472          10.1.3.241:3260 ESTABLISHED
>> tcp        0      0 10.1.6.8:39278          10.1.1.241:3260 ESTABLISHED
>> tcp        0      0 10.1.4.8:41329          10.1.6.241:3260 ESTABLISHED
>> tcp        0      0 10.1.5.8:33553          10.1.5.241:3260 ESTABLISHED
>> tcp        0      0 10.1.3.8:48950          10.1.6.241:3260 ESTABLISHED
>> tcp        0      0 10.1.4.8:54752          10.1.1.241:3260 ESTABLISHED
>> tcp        0      0 10.1.1.8:40911          10.1.4.241:3260 ESTABLISHED
>> tcp        0      0 10.1.4.8:41135          10.1.3.241:3260 ESTABLISHED
>> tcp        0      0 10.1.3.8:44606          10.1.4.241:3260 ESTABLISHED
>> tcp        0      0 10.1.1.8:54677          10.1.3.241:3260 ESTABLISHED
>>
>> (10.1.*.8 is the node and 10.1.*.241 is the SAN)
>>
>> Here is my /etc/multipath.conf:
>>
>> defaults {
>>         path_grouping_policy    multibus
>>         path_selector           readsector0
>>         polling_interval        3
>>         path_selector           "round-robin 0"
>>         failback                immediate
>>         features                "0"
>>         no_path_retry           1
>>         rr_weight               uniform
>>         rr_min_io               100
>> #       user_friendly_names     yes
>> }
>>
>> Both servers are running Ubuntu 12.04LTS.
>>
>> Any ideas?
>>
>> Thanks,
>> Eric
>>
>> -- 
>> dm-devel mailing list
>> dm-devel@redhat.com
>> https://www.redhat.com/mailman/listinfo/dm-devel
>
> -- 
> dm-devel mailing list
> dm-devel@redhat.com
> https://www.redhat.com/mailman/listinfo/dm-devel

  reply	other threads:[~2014-03-25 13:59 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-03-23  4:08 Multipath not using multiple NICs at once Eric
2014-03-23 13:52 ` urgrue
2014-03-25 13:59   ` Eric [this message]
2014-03-23 14:26 ` Hannes Reinecke
2014-03-25 13:55   ` Eric
  -- strict thread matches above, loose matches on Subject: below --
2014-03-22  7:03 Eric

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=53318BD0.7090602@gmail.com \
    --to=iggiggitoo@gmail.com \
    --cc=dm-devel@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body. 
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.