On Wed, Jul 16, 2003 at 10:36:58AM -0500, Edgar Friendly wrote:
> [EMAIL PROTECTED] writes:
>
> >
> > Yeah, that's what I'm thinking and I've gone with just ignoring those
> > DataChunks.
> > >
>
> The thing is, Restarted messages after the data has started streaming
> through should be almost
On Wed, Jul 16, 2003 at 10:36:58AM -0500, Edgar Friendly wrote:
> [EMAIL PROTECTED] writes:
>
> >
> > Yeah, that's what I'm thinking and I've gone with just ignoring those
> > DataChunks.
> > >
>
> The thing is, Restarted messages after the data has started streaming
> through should be almost
[EMAIL PROTECTED] writes:
>
> Yeah, that's what I'm thinking and I've gone with just ignoring those
> DataChunks.
> >
The thing is, Restarted messages after the data has started streaming
through should be almost non-existent. If they happen with any
regularity, it's likely an indication of a
On Tue, Jul 15, 2003 at 09:37:11AM -0500, Edgar Friendly wrote:
> > [Data=... Length=4000 Type=DataChunk time=1f217]
> > [Timeout=23db8 Type=Restarted time=1fd6f]
> > [Data=... Length=4000 Type=DataChunk time=1fd70]
> >
> > Hmm, 1ms between Restarted and the bogus DataChunk. Does that help?
> >
[EMAIL PROTECTED] writes:
> On Mon, Jul 14, 2003 at 10:21:57AM -0600, [EMAIL PROTECTED] wrote:
> > Thanks. This is what I thought the docs were trying to say, but I got the
> > occasional stray DataChunk after Restarted. As you suggested in IRC, I'll
> > treat this as a fatal (for the connection,
On Mon, Jul 14, 2003 at 10:21:57AM -0600, [EMAIL PROTECTED] wrote:
> Thanks. This is what I thought the docs were trying to say, but I got the
> occasional stray DataChunk after Restarted. As you suggested in IRC, I'll
> treat this as a fatal (for the connection, anyway) error and dump a log of
> r
On Mon, Jul 14, 2003 at 10:57:16AM -0500, Edgar Friendly wrote:
> The right way to handle 'Restarted' messages is to go to exactly the
> same state as when you *just* sent your request; i.e. to drop all data
> recieved since then and start waiting for a DataFound message again.
> Note: it's possibl
Bill Currie <[EMAIL PROTECTED]> writes:
> I'm having a bit of a problem with my own FCP client (written in python)
> running into what are either protocol bogosities or misunderstandings (the FCP
> documentation isn't briliantly clear).
>
Assume the protocol was written to be as easy to implement
I'm having a bit of a problem with my own FCP client (written in python)
running into what are either protocol bogosities or misunderstandings (the FCP
documentation isn't briliantly clear).
The following 6 messages form the sequence of events leading up to my
program (really thread) crashing due
On Wednesday 30 October 2002 14:13, you wrote:
> On Wednesday 30 October 2002 14:00 pm, Jay Oliveri wrote:
> > At what point does Fred block during an insert request?
>
> err during an FEC insert.
There isn't really such a thing as a FEC Insert. The FCP FEC commands just
do encoding/decoding
On Wednesday 30 October 2002 14:13, you wrote:
> On Wednesday 30 October 2002 14:00 pm, Jay Oliveri wrote:
> > At what point does Fred block during an insert request?
>
> err during an FEC insert.
There isn't really such a thing as a FEC Insert. The FCP FEC commands just
do encoding/decoding
On Wednesday 30 October 2002 14:00 pm, Jay Oliveri wrote:
> At what point does Fred block during an insert request?
... err during an FEC insert.
--
Jay Oliveri "In the land of the blind,
GnuPG ID: 0x5AA5DD54 the one-eyed man is king."
Ma
At what point does Fred block during an insert request?
I want to know when this happens so I can spawn a new thread at
precisely this moment, and not too soon before it. I suspect the
stable tree spawns a little too early for complexity's sake.
--
Jay Oliveri
On Wednesday 30 October 2002 14:00 pm, Jay Oliveri wrote:
> At what point does Fred block during an insert request?
... err during an FEC insert.
--
Jay Oliveri "In the land of the blind,
GnuPG ID: 0x5AA5DD54 the one-eyed man is king."
Ma
At what point does Fred block during an insert request?
I want to know when this happens so I can spawn a new thread at
precisely this moment, and not too soon before it. I suspect the
stable tree spawns a little too early for complexity's sake.
--
Jay Oliveri
15 matches
Mail list logo