[hylafax-users] Re: So many problems that it's over my head

[David Woolley <david@djwhome.demon.co.uk>]

> 
> 1.   In /var/log/messages i often get the message :
>      Filesystem has SysV-style file creation semantics
>      Does this means anything ?

See the comments around line 365 of hfaxd/FileSystem.c++

>      Can it be disabled ?

Delete the call to logInfo, or don't run with such aggressive system 
logging.  These messages are generated with a priority that is two levels
below anything you should worry about and only one level above debugging
levels.

> 2.   We each day gets a few of :
>      Error: sendfax failed FID: 20000608130629.58865887.obelix

This doesn't seem to originate from Hylafax itself, unless it is new in
the betas.

>      Info: 425 Cannot create data socket (131.166.100.64,4558): Address already
> in use.'

This is almost certainly due to an OS bug.  Linux dynamically allocates port
numbers from a range reserved for server use.  Raise this one with Red Hat,
pointing them at a copy of the assigned numbers RFC.

> 3.   Now and then faxes either gets switched around or one of them
>      gets to the wrong destination (always to another customer)
>      My guess is that the line never gets ATH'ed or similar.

If the system relies on ATH, it is misconfigured.  Modems should be hung
up by dropping DTR, not by the use of ATH, except as a fallback.  In any
case, you should obtain session logs for this one.  You should also consider
your telephone network suspect.

> 4.   We get all sorts of errors often to the same customers,
>      please help us solve the cause, the funny thing is that we
>      could fax to all of them using our old OS/2 fax solution.
>      (which got out of service because of year 2000)
> 
>      Some gets empty pages, some gets condens pages

Condensed pages either result from receivers who don't realise that 
normal resolution faxes have an assymetric resolution, and have a poor
quality TIFF viewer, or from a sending system that has misconfigured
flow control.  Misconfigured flow control will also cause other 
instabilities.

> *1 "Modem protocol error (unknown post-page response)"

As it says.  Hylafax received an invalid response from the remote end/modem.

> *2 "Unknown problem (check modem power)"

See log for details, but probably no response at all from the modems.

> *3 "Failure to train at 2400 bps or +FMINSP value"

Line quality is too poor.  (Or one side or the other is requesting a
minimum speed which is faster than the maximum for the modem pair.)

> *4 "No response to EOP repeated 3 times"

Remote modem disconnected prematurely.

> *5 "Unknown problem (check modem power)"

Duplicate.

> *6 "No carrier detected"

Answered by voice, or confused by automatic fax switch, etc.

No response to EOP is the only one where I think one could even suspect
Hylafax, although its problem might be acting strictly correctly for 
protocol violations, rather than pragmatically.

> 5.   Sometimes it seems that a que is build up and no faxes gets sends
>      It seems that each and every one of them fails but no error messages
>      is given. I am talking about that it sends each fax until maxcount is
>      reached.

Provide the session logs.  This description seems to contradict itself.

Generally, to get help, you need to attack one problem at a time and document
it with a system log, configured for a high level of detail, together with
anything non-standard in your configuration.


____________________ HylaFAX(tm) Users Mailing List _______________________
 To unsub: mail -s unsubscribe hylafax-users-request@hylafax.org < /dev/null