Re: [Nbd] nbd-client with uClibc?
- To: Tom Lloyd <napalmllama@...17...>
- Cc: nbd-general@lists.sourceforge.net
- Subject: Re: [Nbd] nbd-client with uClibc?
- From: Wouter Verhelst <w@...112...>
- Date: Mon, 4 Jun 2007 03:26:48 +0200
- Message-id: <20070604012648.GA8131@...3...>
- In-reply-to: <b2eb75e10706031136y50fd72am1581f51eb261408@...18...>
- References: <b2eb75e10705301050s3f3ad294u3c1ae9ac08d9cf9c@...18...> <20070531002350.GE19644@...3...> <b2eb75e10706021504y7ecf5597oe1821ce7f690ac28@...18...> <200706022329.25775.vapier@...103...> <b2eb75e10706030457l26e652d2y50e7a3fea914851b@...18...> <20070603171552.GA14729@...3...> <b2eb75e10706031136y50fd72am1581f51eb261408@...18...>
On Sun, Jun 03, 2007 at 07:36:42PM +0100, Tom Lloyd wrote:
> Yup, it compiles now. Thanks!
>
> I've got another problem, though - it seems to drop the connection
> after a short while with the following message:
>
> muttley ~ # nbd-client 10.0.0.2 555 /dev/nbd0
> Negotiation: ..size = 10485760KB
> bs=1024, sz=10485760
> muttley ~ # Kernel call returned: Broken pipeClosing: que, sock, done
>
> Is this normal?
Okay. You're not the first to hit this, and probably not the last
either. I therefore reverted r231, and released 2.9.3. That should fix
this. I'll do the same on the 2.8 branch later, at an hour that does not
read as '3:26 AM local time'.
(r231 implements non-blocking sockets for communication with the client.
That improved performance, but apparently it isn't bug-free... I'll
leave this as is now, except on trunk, where I'll intend to fix the
issue rather than leave it)
--
Shaw's Principle:
Build a system that even a fool can use, and only a fool will
want to use it.
Reply to: