Re: [hylafax-users] Fax stopped working

[Lee Howard <faxguy@howardsilvan.com>]

On 2003.07.16 19:15 Ross Boylan wrote:
> Although I have faxed successfully from my machine in the past
> (including using HylaFax), recently things have not worked.  It's
> probably been a few months since I last tried.
> 
> I've tried at least 3 different fax numbers, but I seem to keep
> getting errors like
> No response to EOP repeated 3 times  or
> No response to MPS repeated 3 times
> 
> In my last attempt, the recipient told me she got the cover page 3
> times, but nothing more.  I don't know if the whole cover made it.
> That job failed with an MPS error.
> 
> While my software environment has been changing, I don't think my
> HylaFax server has.  Nor have we done anything with our phone lines,
> though they have always been a little marginal.
> 
> If there is a way to fix or diagnose this problem further, I'd like to
> try.  I've been working on some other fax problems (client side).  Now
> that I think I've fixed them, it's disappointing to find I can't fax
> at all!  On the other hand, I have a sinking feeling it might be a
> hardware problem, though I don't know if it's the line or the card.
> 
> Thanks for any help.
> 
> Some details follow:
> -------------------------------------------------
> My modem is U.S. Robotics 56K Voice INT, model 5685. I don't
> know how to check the firmware version but it's an ISA PNP modem
> that's
> 2-3 years old.  It's not a Courier or a Sporster, but it was supposed
> to be good quality.  I've noticed the FAQ dissing the quality of US
> Robotics modems, but it's what I've got.
> 
> Hylafax-server 4.1.1-1.1 on a Debian/testing system with AMD
> processor.  It has lots of peripherals and interrupts are scarce.
> This version appears to be after the changes identified in
> http://bugs.hylafax.org/bugzilla/show_bug.cgi?id=185, which sounded
> related.
> 
> My last failure log:
> Jul 16 16:56:50.31: [ 7119]: SESSION BEGIN 00000058 14154766153
> Jul 16 16:56:50.31: [ 7119]: HylaFAX (tm) Version 4.1.1
> Jul 16 16:56:50.31: [ 7119]: SEND FAX: JOB 59 DEST 476-6153 COMMID
> 00000058 DEVICE '/dev/ttyS2'
> Jul 16 16:56:58.48: [ 7119]: STATE CHANGE: RUNNING -> SENDING
> Jul 16 16:58:03.47: [ 7119]: SEND: No response to MPS repeated 3
> times; Giving up after 3 attempts to send same page
> "docq/doc59.cover;31", dirnum 0
> Jul 16 16:58:03.58: [ 7119]: STATE CHANGE: SENDING -> MODEMWAIT
> (timeout 5)
> Jul 16 16:58:03.58: [ 7119]: SESSION END

Please set SessionTracing to 0xFFF for a more detailed log; this one 
doesn't help much.

I can't say for certain that upgrading is going to resolve the problem, 
but I certainly am going to recommend it.  There have been a few 
security issues, numerous bugfixes, and some protocol fixes which 
you'll find present in 4.1.6 and/or current CVS (current CVS doesn't 
disable pagechopping).

Also, if you're using Class 2.0 I would highly recommend that you 
switch to Class 1.  USR's firmware seems perpetually flawed, and even 
if they do quietly fix it without any communication with the customer 
base, there certainly is no way to update the firmware on most of their 
modems (the Couriers being the usual exception there).

Lastly, I would like to say a word or two about USR modems and this "no 
response to MPS or EOP" error...

I've used many, many USR modems in my time with HylaFAX.  In fact, I 
started using HylaFAX with a USR 3CP5610 a few years ago in a 
development system.  Later, I ended up using a USR clone (ISA version 
using the same TI chipset USR used on their Sportsters) in a production 
server.  After finally giving up on USR's Class 2.0 and switching to 
Class 1 I still got too many "no response to MPS or EOP" errors.  Since 
that time I've done a lot of work trying to figure out what it is about 
USR modems that makes this error so common.  I've looked into timing 
matters, inserting pauses here and there, determining the threshold for 
those pauses, etc.  I've also tried to determine if it is related to 
certain data compression types or other negotiation parameters.  Also I 
tried to determine if it was related to the type (brand, model) of 
remote system.  There have been some signifiant improvements along the 
way, most of which have been in the form some new configuration 
parameters in HylaFAX, but ultimately I still get more "no response to 
MPS or EOP" errors than can be blamed on either a line disconnection or 
some error on the remote end.  These days I have two USRs in production 
use, an old Courier and a 3CP5610C, and I have an old 1996 ISA-jumpered 
Sportster in a development system (to test HylaFAX on the most flawed 
firmware I can find).  I know how flawed the old ISA one is because I 
bang on it in my development system, but I don't know how badly it 
would perform in production use.  The 3CP5610C works better in 
production use than does the old Courier, but I still get an occasional 
"no response to MPS or EOP" error.

On the remote side some say they get blank pages, others say they get 
partial pages, and others say they get no page at all.  All of these 
may be fascets of the same error, only varying by how the remote system 
handles it.  Some systems may not disconnect quickly and so we may 
actually be able to renegotiate and resend the page, in which case the 
remote may get it three times.  The "partial page" is what disturbs 
me.  The fact that the remote system started getting data at all 
indicates that our protocol up until then was fine, and it indicates 
the exact point of failure: during the image data transfer.  In my 
estimation that's a bad sign - the image data transfer function doesn't 
change mid-way through a page.  So, it was working and then stopped 
working, and I can't see how that can be a result of anything but 
hardware failure, especially when it often occurs with 0ms/scanline 
sessions (where the remote doesn't require short "catch-up" pauses 
after each line of image data).

In any case... whenever we find a hardware/firmware flaw in a modem 
that cannot have the firmware upgraded (either because it's not 
upgradable or its out of production and support) then we have to look 
at a new modem to resolve the matter.  I really am quite pleased with 
MultiTech's MT5634ZBA-V92 modems.  There have been bugs, but most of 
them have had easy workarounds, and MultiTech does readily service fax 
support incidents and eventually releases updated firmware to fix the 
problem, and it can usually be "flashed" directly on the HylaFAX box.

Anyway, that's not really a sales-pitch so much as it is my experience.

Lee.

____________________ HylaFAX(tm) Users Mailing List _______________________
  To subscribe/unsubscribe, click http://lists.hylafax.org/cgi-bin/lsg2.cgi
 On UNIX: mail -s unsubscribe hylafax-users-request@hylafax.org < /dev/null
  *To learn about commercial HylaFAX(tm) support, mail sales@hylafax.org.*