Re: write_sock stream errors

[email protected] (Peter T. Breuer)
Newsgroups gmane.linux.enbd.general
Message-ID <[email protected]>
Nikos Gkorogiannis <[email protected]> wrote:
> I'm running 2.4.31 on 2.6.7 (server and client) and the server, very
> shortly after connection with the client, gives these out:
> 
> slave1 enbd-server: stream   548: <# 585> write_sock stream 0x8065734
> has closed socket  
> slave1 enbd-server: enbd-server   548: <#1957> talk_hello Introduction
> failed: Invalid argument  
> slave1 enbd-server: stream   551: <# 585> write_sock stream 0x8065734
> has closed socket  
> slave1 enbd-server: enbd-server   551: <#1957> talk_hello Introduction
> failed: Invalid argument  

Server with pids 548 and 551 complained about a closed socket when they
tried to write. Line 1957 of enbd-server.c. Seems to be in
negotiate(). 

I think the servers ought to die (?) and restart the negotiation.  You
can check the pids. It may be a race condition. It's the first write on
the newly opened stream through the session port. The write attempt
returned a negative number and errno=EINVAL was set. I know no more.
Man page for "write" says:

  EINVAL fd is attached to an object which is unsuitable for writing.


> slave1 enbd-server: enbd-server: server (0) set default signal handlers
> for slave server 548  
> slave1 enbd-server: enbd-server: server (0) set default signal handlers
> for slave server 551  
> 
> The client does not get any errors (at least under the IO load I'm
> subjecting it). Here's nbdinfo:

Well, maybe the protocol of having the server write the hello first is
a bit silly :-).

> It's not necessarily a show-stopper, as there are no errors generated on
> the client, but could it be? And what do they mean?

Could it be that the server is faster than the client?

Peter
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.