Hello,
For sometime I am discussing my problem on hylafax list. In short, I most probably have a problem with serial driver which is "asy" in my system. My system is receiving faxes thru MultiTech ZDX external modem. During the day, I have about 10-15 modem "wedge" problems. Only solution is a reboot of the system which then leads to worse problems.
System is doing several jobs with the help of zones. I have all patches applied in my system. Several Hylafax versions are tried and current one is the latest version 5.2.7, and is compiled in my system. Several suggestions of hylafax list are checked, unfortunately none helped me to solve my problem.
I am told from hylafax list that there may be a bug in Solaris serial driver. Can anybody shed me a light please?
Reverting the system back to FreeBSD is definitly not appreciated on my side. However, this problem will definetly be a big trouble if I can not solve it soon.
Current eeprom output in my system is as follows:
bash-3.00# eeprom
ata-dma-enabled=1
atapi-cd-dma-enabled=0
ttyb-rts-dtr-off=false
ttyb-ignore-cd=true
ttya-rts-dtr-off=false
ttya-ignore-cd=false
ttyb-mode=9600,8,n,1,-
ttya-mode=19200,8,n,1,s
lba-access-ok=1
prealloc-chunk-size=0x2000
bootpath=/pci@0,0/pci-ide@1f,2/ide@0/cmdk@0,0:a
keyboard-layout=TurkishQ
console=text
bash-3.00#
I do not have anything put in my /etc/systems file for this problem. My trial of
asy:default_dtrlow=0
did not help. There is no such tunable in asy driver error message is taken during system startup.
Below is slightly long historical information about the problem:
---
De: hylafax-users-bounce@hylafax.org
mailto:hylafax-users-bounce@hylafax.org En nombre de Ertan Küçükoglu
Enviado el: miércoles, 27 de agosto de 2008 11:40
Para: hylafax-users@hylafax.org
Asunto: [hylafax-users] Solaris x86 "T.30 T2 timeout, expected page not received" or "Missing EOL after 5 seconds" problem
Hello,
I have frequent hang of modem after switching to Solaris 10 Update 5 (7/08) x86 OS.
I recently switched my system from FreeBSD 7.0 to Solaris x86. All patches are applied to Solaris system. After this switch I got a lot of trouble because of hanging of the modem. My old MultiTech 56K ZDX modem started to freeze for some jobs. Below is a sample log from one of them. All available classes are tried until now. Modem is set as Class1 as this is the long lasting class before freezing.
What I observed is:
First below message comes from system. Then after 5 minutes I get “modem on /dev/cua0 appears wedged” e-mails from system. I manually power recycle the modem. Kill running faxgetty process. There is no lights blinking on the modem. Server cannot initialize the modem. Only solution on my side is to reboot the system. This reboot is a real big problem for me. This box is doing mail server, minor database server, fax server, web proxy jobs. Company connection to world is blocked for 2-5 minutes.
Also, when I setup modem for Class2 I read below lines in the log file, and I need to restart the server again like in Class1.
Aug 21 17:49:33.46: [ 607]: RECV: Missing EOL after 5 seconds
Aug 21 17:49:33.46: [ 607]: RECV: end page
I am using source compiled version of Hylafax 4.4.4.
Any help will be appreciated.
My log files are as follows:
recvq/fax000000165.tif (ftp://ana.ozlerplastik.com:4559/recvq/fax000000165.tif):
Sender: <UNSPECIFIED>
Pages: 2
Quality: 100 lines/inch
Size: ISO A4
Received: 2008:08:26 11:50:18
Time To Receive: 0:37
Signal Rate: 14400 bit/s
Data Format: 2-D MR
Error Correct: No
CallID1:
CallID2:
Received On: cua0
CommID: 000000139 (ftp://ana.ozlerplastik.com:4559/log/c000000139)
bash-3.00# cat c000000139
Aug 26 11:50:03.73: [ 598]: SESSION BEGIN 000000139 902124566697
Aug 26 11:50:03.73: [ 598]: HylaFAX (tm) Version 4.4.4
Aug 26 11:50:03.73: [ 598]: <-- [4:ATA\r]
Aug 26 11:50:09.04: [ 598]: --> [7:CONNECT]
Aug 26 11:50:09.04: [ 598]: ANSWER: FAX CONNECTION DEVICE '/dev/cua0'
Aug 26 11:50:09.04: [ 598]: RECV FAX: begin
Aug 26 11:50:09.04: [ 598]: <-- data [32]
Aug 26 11:50:09.04: [ 598]: <-- data [2]
Aug 26 11:50:11.06: [ 598]: --> [7:CONNECT]
Aug 26 11:50:11.06: [ 598]: <-- data [23]
Aug 26 11:50:11.06: [ 598]: <-- data [2]
Aug 26 11:50:11.89: [ 598]: --> [7:CONNECT]
Aug 26 11:50:11.89: [ 598]: <-- data [13]
Aug 26 11:50:11.89: [ 598]: <-- data [2]
Aug 26 11:50:12.54: [ 598]: --> [2:OK]
Aug 26 11:50:12.54: [ 598]: <-- [9:AT+FRH=3\r]
Aug 26 11:50:13.14: [ 598]: --> [7:CONNECT]
Aug 26 11:50:14.11: [ 598]: --> [2:OK]
Aug 26 11:50:14.11: [ 598]: REMOTE wants 14400 bit/s
Aug 26 11:50:14.11: [ 598]: REMOTE wants A4 page width (215 mm)
Aug 26 11:50:14.11: [ 598]: REMOTE wants unlimited page length
Aug 26 11:50:14.11: [ 598]: REMOTE wants 200 x 100 dpi
Aug 26 11:50:14.11: [ 598]: REMOTE wants 2-D MR
Aug 26 11:50:14.11: [ 598]: RECV recv DCS (command signal)
Aug 26 11:50:14.11: [ 598]: RECV training at v.17 14400 bit/s
Aug 26 11:50:14.11: [ 598]: <-- [11:AT+FRM=145\r]
Aug 26 11:50:15.85: [ 598]: --> [7:CONNECT]
Aug 26 11:50:17.41: [ 598]: RECV: TCF 2776 bytes, 1% non-zero, 2729 zero-run
Aug 26 11:50:17.41: [ 598]: --> [10:NO CARRIER]
Aug 26 11:50:17.41: [ 598]: <-- [9:AT+FRS=7\r]
Aug 26 11:50:17.50: [ 598]: --> [2:OK]
Aug 26 11:50:17.50: [ 598]: TRAINING succeeded
Aug 26 11:50:17.50: [ 598]: <-- [9:AT+FTH=3\r]
Aug 26 11:50:17.55: [ 598]: --> [7:CONNECT]
Aug 26 11:50:17.55: [ 598]: <-- data [3]
Aug 26 11:50:17.55: [ 598]: <-- data [2]
Aug 26 11:50:18.86: [ 598]: --> [2:OK]
Aug 26 11:50:18.86: [ 598]: <-- [11:AT+FRM=146\r]
Aug 26 11:50:19.74: [ 598]: --> [7:CONNECT]
Aug 26 11:50:19.74: [ 598]: RECV: begin page
Aug 26 11:50:46.25: [ 598]: RECV: 1149 total lines, 0 bad lines, 0 consecutive bad lines
Aug 26 11:50:46.25: [ 598]: RECV: end page
Aug 26 11:50:46.25: [ 598]: --> [10:NO CARRIER]
Aug 26 11:50:46.25: [ 598]: <-- [9:AT+FRH=3\r]
Aug 26 11:50:46.67: [ 598]: --> [7:CONNECT]
Aug 26 11:50:47.39: [ 598]: --> [2:OK]
Aug 26 11:50:47.39: [ 598]: RECV recv MPS (more pages, same document)
Aug 26 11:50:47.39: [ 598]: <-- [9:AT+FRS=7\r]
Aug 26 11:50:47.57: [ 598]: --> [2:OK]
Aug 26 11:50:47.57: [ 598]: <-- [9:AT+FTH=3\r]
Aug 26 11:50:47.71: [ 598]: --> [7:CONNECT]
Aug 26 11:50:47.71: [ 598]: RECV send MCF (message confirmation)
Aug 26 11:50:47.71: [ 598]: <-- data [3]
Aug 26 11:50:47.71: [ 598]: <-- data [2]
Aug 26 11:50:47.71: [ 598]: RECV FAX (000000139): from <UNSPECIFIED>, page 1 in 0:37, INF, 200 x 100 dpi, 2-D MR, 14400 bit/s
Aug 26 11:50:49.03: [ 598]: --> [2:OK]
Aug 26 11:50:49.03: [ 598]: <-- [11:AT+FRM=146\r]
Aug 26 11:50:49.82: [ 598]: --> [7:CONNECT]
Aug 26 11:50:49.82: [ 598]: RECV: begin page
Aug 26 11:52:35.04: [ 598]: <-- data [1]
Aug 26 11:52:35.04: [ 598]: RECV: Missing EOL after 5 seconds
Aug 26 11:52:35.04: [ 598]: RECV: end page
Aug 26 11:52:40.05: [ 598]: MODEM <Timeout>
Aug 26 11:52:40.05: [ 598]: <-- data [1]
Aug 26 11:52:40.26: [ 598]: MODEM <Timeout>
Aug 26 11:52:40.26: [ 598]: RECV keeping unconfirmed page
Aug 26 11:52:40.26: [ 598]: RECV FAX: T.30 T2 timeout, expected page not received
Aug 26 11:52:40.26: [ 598]: <-- [9:AT+FTH=3\r]
Aug 26 11:52:40.26: [ 598]: RECV FAX (000000139): from <UNSPECIFIED>, page 2 in 0:37, INF, 200 x 100 dpi, 2-D MR, 14400 bit/s
Aug 26 11:52:40.26: [ 598]: RECV FAX (000000139): recvq/fax000000165.tif from <UNSPECIFIED>, route to <unspecified>, 2 pages in 2:31
Aug 26 11:52:47.82: [ 598]: --> [0:]
Aug 26 11:52:47.82: [ 598]: RECV FAX (000000139): session with <UNSPECIFIED> terminated abnormally: T.30 T2 timeout, expected page not received
Aug 26 11:52:47.82: [ 598]: RECV FAX: bin/faxrcvd "recvq/fax000000165.tif" "cua0" "000000139" "T.30 T2 timeout, expected page not received"
Aug 26 11:52:47.82: [ 598]: RECV FAX: end
Aug 26 11:52:47.82: [ 598]: T.30 T2 timeout, expected page not received
Aug 26 11:52:47.82: [ 598]: SESSION END
bash-3.00#
---
Below is what I have received from the hylafax list
---
-----Original Message-----
From: hylafax-users-bounce@hylafax.org [mailto:hylafax-users-
bounce@hylafax.org] On Behalf Of Lee Howard
Sent: Thursday, August 28, 2008 6:17 AM
To: Ertan Küçükoğlu
Cc: hylafax-users@hylafax.org
Subject: Re: [hylafax-users] Solaris x86 "T.30 T2 timeout, expected
page not received" or "Missing EOL after 5 seconds" problem
Ertan Küçükoğlu wrote:
I have frequent hang of modem after switching to Solaris 10 Update 5
(7/08) x86 OS.
I recently switched my system from FreeBSD 7.0 to Solaris x86. All
patches are applied to Solaris system. After this switch I got a lot
of trouble because of hanging of the modem. My old MultiTech 56K ZDX
modem started to freeze for some jobs. Below is a sample log from
one of them. All available classes are tried until now. Modem is set
as
Class1 as this is the long lasting class before freezing.
So this very same hardware was working fine with the very same version
of HylaFAX under FreeBSD before switching to Solaris where the
problemsActually, there is a new hardware in terms of Computer. I have a new HP Proliant ML110 G5 server for Solaris. Unfortunately, I do not have the chance to make a trial with FreeBSD on this hardware soon. Since this is a production server right now. My earliest trial time may be 13th September.
However, I do not hit this error always. Only some systems sending fax to me is not completing properly (as shown in my first e-mail with the subject).
Right after this not completion, modem stays On Hook.
started? If so, it would seem to be either a bug in the Solaris
serial driver or a problem in how HylaFAX handles the serial port on
Solaris (that doesn't always cause the modem to lock up).I have seen DTR delay related pages related to Solaris SPARC versions. It seems that Solaris SPACR serial driver used to have default 3 seconds delay on DTR. I am in progress to find related driver and set parameter to pass to the kernel in order to reduce this delay for Solaris X86 version. However, this delay problem seems to be a problem long ago. Anyway, I am checking if this is still an issue.
If you power-cycle the modem (it's external, after all) does it wake
up? I suspect not, which would mean that the problem is in the serial
interfacing.If I do not reboot the Solaris system, only power-cycle modem which is an external one problem continues. Only solution is to reboot, or power-cycle the Solaris system.
There is an unlikely possibility that the slightly different modem
reset procedure (which only comes into play in potential wedged
conditions) in
HylaFAX+ would be able to recover modem connectivity. You may want to
give that a try. Assuming that does not work you'll need to getIs it possible to learn this procedure? I do not know how to do this.
debugging information out of the Solaris kernel serial driver in order
to proceed with any predictability.
Thanks,
Lee.Regards,
Ertan
---
Finally, What I found in "asy" driver manual page also did not help me to solve my problems:
---
Ok, I made some search and found following.
1- Solaris for x86 platform uses asy driver for COM ports.
bash-3.00# ls -l /dev/cua0
lrwxrwxrwx 1 root root 29 Aug 9 15:25 /dev/cua0 ->
../devices/isa/asy@1,3f8:a,cu
bash-3.00#
2- asy driver manual page have following entry in it:
Note - This module is affected by the setting of certain
eeprom variables, ttya-ignore-cd and ttya-rts-dtr-
off (and similarly for ttyb-, ttyc-, and ttyd-
parameters). For information on these parameters,
see the eeprom(1M) man page.
Note - For serial ports on the standard COM1 to COM4 I/O
addresses above, the default setting for ttya-
ignore-cd and ttya-rts-dtr-off is true. If any of
these ports are connected to a modem, these settings
should be changed to false. For serial ports on
non-standard I/O addresses, the default setting for
ttya-ignore-cd and ttya-rts-dtr-off is false.
What I did now is as follows:
bash-3.00# eeprom ttya-ignore-cd
ttya-ignore-cd=true
bash-3.00# eeprom ttya-ignore-cd=false
bash-3.00# eeprom ttya-ignore-cd
ttya-ignore-cd=false
bash-3.00# eeprom ttya-rts-dtr-off
ttya-rts-dtr-off=false
bash-3.00# reboot
Since ttya-rts-dtr-off was false. I did not touch it. I will observe the behavior for next week and provide a feedback to the list.
Regards,
Ertan
Edited by: kucukoglu on Sep 5, 2008 10:40 PM