lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [day] [month] [year] [list]
Message-ID: <20210503073849.27f9b4c0@hermes.local>
Date:   Mon, 3 May 2021 07:38:49 -0700
From:   Stephen Hemminger <stephen@...workplumber.org>
To:     netdev@...r.kernel.org
Subject: Fw: [Bug 212921] New: ECMP not working for local sockets



Begin forwarded message:

Date: Sun, 02 May 2021 07:10:44 +0000
From: bugzilla-daemon@...zilla.kernel.org
To: stephen@...workplumber.org
Subject: [Bug 212921] New: ECMP not working for local sockets


https://bugzilla.kernel.org/show_bug.cgi?id=212921

            Bug ID: 212921
           Summary: ECMP not working for local sockets
           Product: Networking
           Version: 2.5
    Kernel Version: 5.8.0-50-generic
          Hardware: All
                OS: Linux
              Tree: Mainline
            Status: NEW
          Severity: normal
          Priority: P1
         Component: IPV4
          Assignee: stephen@...workplumber.org
          Reporter: nitin.i.joy@...il.com
        Regression: No

When you're creating local TCP sockets in a Linux machine, the connections to
the same destination IP are not load-balanced across multiple interfaces when
ECMP path is set. Even when net.ipv4.fib_multipath_hash_policy is set to L4
hash, multiple interfaces are never used for same destination. 

I tried working around the issue by setting two route table entries with same
metric using `ip route append` command. In this case, the connections get
load-balanced across multiple interfaces for 5-10 seconds, after which all
future connections will choose one of the interfaces. There is no configuration
that can disable this behavior. I tried disabling tcp_metrics_nosave

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are the assignee for the bug.

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ