[Bugs] [Bug 1409189] Failed to set TCP_USER_TIMEOUT msgs seen in logs

[bugzilla at redhat.com]

https://bugzilla.redhat.com/show_bug.cgi?id=1409189

SATHEESARAN <sasundar at redhat.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sasundar at redhat.com



--- Comment #1 from SATHEESARAN <sasundar at redhat.com> ---
I had a chat regarding this issue with Niels and I am feeling bad that I
haven't followed up. Just pasting that mail conversation, so that it could help

On Wed, Jul 22, 2015 at 11:49:56AM +0530, SATHEESARAN wrote:
> Hi Niels,
>
> I have observed warning messages in glusterd related to TCP_USER_TIMEOUT :
>
> <snip>
> [2015-07-22 11:38:41.979511] I [rpc-clnt.c:972:rpc_clnt_connection_init]
> 0-management: setting frame-timeout to 600
> [2015-07-22 11:38:41.987058] W [socket.c:923:__socket_keepalive] 0-socket:
> failed to set TCP_USER_TIMEOUT -1000 on socket 8, Invalid argument
> [2015-07-22 11:38:41.987099] E [socket.c:3018:socket_connect] 0-management:
> Failed to set keep-alive: Invalid argument
> </snip>
>
> Do you know what these messages mean ?


That means that __socket_keepalive() is called with a timeout (last
parameter) of -1. This value normally comes from a socket_private_t
structure and is set through "transport.tcp-user-timeout".

Most functions expect the timeout be an unsigned int, how it gets to a
-1 value is unclear to me.

Does this happen always, or only in some configuration/environment?

Thanks,
Niels

-- 
You are receiving this mail because:
You are on the CC list for the bug.
You are the assignee for the bug.