provided by Star2Billing S.L.

Hello people !

When Asterisk is starting execution of a2billing.php is taking very-very slow to arive to playback of prompt messages.
This delay is about 12 sec. if I am running with debug off, and up to 20 sec with full debug.

In simple words,the whole processing DeadAGI script takes too long time ...Is it normal ???

I see no significant errors in A2billing log,nor Asterisk console.

By the way,I replaced with Open_A2billing_Racoon from Trixbox distribution to see what happens.And then delay decreases almost as twice. But is still no acceptable to listen on the phone silence a 5 seconds.

I think problem hides withing PHPAGI on PHP5.Maybe..Dont know how to find that.

I have installed latest Asterisk2Billing V1.2.3.
this configuration -->
Opensuse 10.1 : PIII/256 RAM
PHP 5.1.2 / MySQL 5.0.18
Asterisk 1.2.5

I attached debug here.

Wild guess: such a long delay could be caused by disk I/O. Like a disk w/ bad sectors or a nfs drive which doesn't respond. What do other system logs say?

Also one of the possible reasons:

also Apache i guess...
This all software what like to eat too much the memory ever.
Do you not forgot to create and mount the Swap file, i hope?
show me please answer from this commands:


will see..

I have swap 768 mb
-------------------------------------------------------------------------
fdisk -l

Disk /dev/hda: 40.0 GB, 40007761920 bytes
255 heads, 63 sectors/track, 4864 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

Device Boot Start End Blocks Id System
/dev/hda1 * 1 730 5863693+ b W95 FAT32
/dev/hda2 731 4864 33206355 f W95 Ext'd (LBA)
/dev/hda5 731 2010 10281568+ 83 Linux
/dev/hda6 2011 2108 787153+ 82 Linux swap / Solaris
/dev/hda7 2109 4864 22137538+ 83 Linux
--------------------------------------------------------------------------
cat /etc/fstab

/dev/hda5 / ext3 acl,user_xattr 1 1
/dev/hda7 /home ext3 acl,user_xattr 1 2
/dev/hda6 swap swap defaults 0 0
proc /proc proc defaults 0 0
sysfs /sys sysfs noauto 0 0
debugfs /sys/kernel/debug debugfs noauto 0 0
usbfs /proc/bus/usb usbfs noauto 0 0
devpts /dev/pts devpts mode=0620,gid=5 0 0
/dev/fd0 /media/floppy auto noauto,user,sync 0 0

/dev/hda1 /media/windows vfat defaults 0 0
-----------------------------------------------------------------------

I found something in /var/log/messages

is a bunch of complains about var: Deprecated in /adodb/adodb.inc.php .

what saying 'top' ?

Added after 4 minutes:

Do you have X (Desktop) installed on this linux machine?

Here is a top output . CPU usage increasing not more than 5% during call activity.

top - 23:43:29 up 7:46, 3 users, load average: 0.27, 0.24, 0.20
Tasks: 93 total, 2 running, 90 sleeping, 0 stopped, 1 zombie
Cpu(s): 2.4% us, 0.3% sy, 0.0% ni, 93.3% id, 3.7% wa, 0.3% hi, 0.0% si
Mem: 248800k total, 214560k used, 34240k free, 18212k buffers
Swap: 787144k total, 95280k used, 691864k free, 101576k cached

I noticed now HDD activity during this delay. Actualy I am thinking is logging errors at that moment.

Yes,I am running X ,Hey - is my only working machine

But I tested all stuff already without X few times in plain console. Same effect -no changes at all.

Slow as a dead pig

show the output please

hdparm -tT /dev/hda

/dev/hda:
Timing cached reads: 552 MB in 2.01 seconds = 275.27 MB/sec
Timing buffered disk reads: 62 MB in 3.05 seconds = 20.30 MB/sec

Disk I have is Toshiba 40GB SuperSlim MK4026GAX Hard Drive.(2.5).

I found some specs
Data Transfer Rate 100 MBps
Buffer 16 Mb
Seek time 2/12/22 ms
Average Rotational Latency 5.55 ms
Rotational Speed 5400 RPM

I made a short HDD test with S.M.A.R.T tools.

I am stopping syslog-ng ,and delay drops to 3 sec.
... continuing to digg...

try to remove the logging and verbose option !
and let us know

A.

Judging from your first /var/log/messages, sth. is tragically wrong with the way your system tries to log things. I could imagine such a thing as 'localhost' not pointing to the loopback i/f, so that each attempt to use a service will result in network timeouts.

Do other things at your machine work? What services/settings does it have?

Well,yesterday I killed completly my root partition,and had to reinstall fresh system,asterisk,freepbx and a2billing .
Same thing ... Fresh Opensuse 10.1..

To Areski


What do you mean by this ? Logging & debug options in a2billing.conf ?

I did already debug=0 ,logger_enable=NO .

I found now more things ,is not only /var/log/messages was full of above mentioned var eprecated from php,but /var/log/warn had same thing.

I looked in syslog-ng filters and found this :

(1) log filter f_messages (var/log/messages)
not facility(news , mail) and not filter(f_iptables)

&

(2) f_warn log filter f_warn (var/log/warn)
level(err , warn , crit) and not filter(f_iptables)

Looks like those are both destinations for all WARN messages

Then when I am changing these records like that :

(1) not facility(news , mail) and not filter(f_iptables) and not level(warn)

(2) level(err , crit) and not filter(f_iptables)

problem is gone Delay drops to 2-3 sec. ,and HDD is sleeping

I added new filter f_warnings pointing to a new log file /var/log/warnings
level(warn)

In this case all errors appearing in new log.

What i think that misconfigured syslog-ng from Suse simple overloads the log files with not needed warning of deprecated functions of PHP.

Am I am right ?

If someone of you could just look out of curiosity at configs of syslog on other than Suse.

Sleeping??
That IS your problem then! The hdd should never be programmed to sleep!
(repeat) You never let your hdd sleep (=spin down) on a server-like system!
Quoting google, each spin-up cycle equals several days of hdd wear-off..

Anyway, greekman, this is not an a2b bug

At least ugly solution....
normaly the code of A2B must be reviewed for compatibility on PHP5 I think.

I am closing my *bug* here now.

Thanks for help
Eyxaristo gia tin boithia