You need to use distro you know and which suits you best.

I've been using Debian in combination with kannel for quite some time
(about 5 years) and never had any problems.

Currently I'm running kannel as well on top of Debian Wheezy it works just
fine.

If you are having troubles with kannel's performance. You need to read
mailing lists a bit. Main bottlenecks were discussed many times, in most
cases it's not kannel's issue.

Instead of general MySQL I'd recommend using MariaDB 5.5 (mariadb.org). It
uses XtraDB for InnoDB engine and is very well maintained. There is debian
repo for it as well.

Combination of max-pending-submits  / throughput should be set accordingly
to your smsc upstream provider. Basically max-pending-submits means a
number of outstanding operations between you and smsc upstream. If you
don't have any errors like throttling or queue full in your smsc log
current values should be fine.

2016-01-13 15:01 GMT+03:00 Otandeka Simon Peter <sotand...@gmail.com>:

> Hi,
>
> Try setting the sms-resend-freq and sms-resend-retry to reduce on the
> number of sms retries. Also set a low figure on the max-pending-submits to
> reduce on your queue. This might help.
>
> But also check on your conxn plus you could be sending to
> non-existent/wrongly formatted numbers, the smsc rejects and your system
> keeps trying to re-send them.
>
> If the issue occurs at a certain period all the time, check with your smsc
> vendor. Sometimes they have queue/throttling problems.
>
> My 2 cents.
>
> @sotandeka
>
>
> On Wed, Jan 13, 2016 at 2:43 PM, Grant Saicom <grant.sai...@gmail.com>
> wrote:
>
>> Example log below of one very common warning. I cannot see anything about
>> tcp retransmission in the logs. We see it in tshark when things start going
>> a bit out of sorts.
>>
>> PDU has been sanitised for privacy, so the destination and text has been
>> removed.
>>
>> 2016-01-13 11:00:00 [4771] [7] DEBUG: SMPP[internal]: Sending PDU:
>> 2016-01-13 11:00:00 [4771] [7] DEBUG: SMPP PDU 0x1c5bfa0 dump:
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   type_name: submit_sm
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   command_id: 4 = 0x00000004
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   command_status: 0 = 0x00000000
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   sequence_number: 6541387 =
>> 0x0063d04b
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   service_type: "P_289361630"
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   source_addr_ton: 2 = 0x00000002
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   source_addr_npi: 1 = 0x00000001
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   source_addr: "435719276"
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   dest_addr_ton: 2 = 0x00000002
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   dest_addr_npi: 1 = 0x00000001
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   destination_addr: ""
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   esm_class: 3 = 0x00000003
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   protocol_id: 0 = 0x00000000
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   priority_flag: 0 = 0x00000000
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   schedule_delivery_time: NULL
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   validity_period: NULL
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   registered_delivery: 1 =
>> 0x00000001
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   replace_if_present_flag: 0 =
>> 0x00000000
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   data_coding: 0 = 0x00000000
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   sm_default_msg_id: 0 = 0x00000000
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   sm_length: 91 = 0x0000005b
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   short_message:
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:    Octet string at 0x1cf1070:
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:      len:  91
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:      size: 92
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:      immutable: 0
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:      data: 59
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:      data: 48
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:      data: 4
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:      data:
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:      data: 49 53
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:      data: 4c 5
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:    Octet string dump ends.
>> 2016-01-13 11:00:00 [4771] [7] DEBUG: SMPP PDU dump ends.
>> 2016-01-13 11:00:00 [4771] [7] WARNING: SMPP: PDU element <service_type>
>> too long (length is 11, should be 5)
>> 2016-01-13 11:00:00 [4771] [7] DEBUG: SMPP[internal]: throughput
>> (12.00,0.00)
>> 2016-01-13 11:00:00 [4771] [7] DEBUG: SMPP[internal]: Sending PDU:
>> 2016-01-13 11:00:00 [4771] [7] DEBUG: SMPP PDU 0x1d5d380 dump:
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   type_name: submit_sm
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   command_id: 4 = 0x00000004
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   command_status: 0 = 0x00000000
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   sequence_number: 6541388 =
>> 0x0063d04c
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   service_type: "P_289340514"
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   source_addr_ton: 2 = 0x00000002
>> 2016-01-13 11:00:00 [4771] [7] DEBUG:   source_addr_npi: 1 = 0x00000001
>>
>> On 13 Jan 2016, at 13:34, Grant Saicom <grant.sai...@gmail.com> wrote:
>>
>> Hi
>>
>> Thank you for your reply. See my config below. Logs are pretty massive,
>> so I am going to try find an example and send through as well.
>>
>>
>> ================================Kannel Conf======================
>> group = core
>> admin-port = 13000
>> smsbox-port = 13001
>> admin-password = ########
>> log-file = "/var/log/kannel/kannel-bearerbox.log"
>> log-level = 0
>> box-deny-ip = "*.*.*.*"
>> box-allow-ip = "127.0.0.1"
>> dlr-storage = mysql
>> store-location = "/var/spool/kannel"
>> smsbox-max-pending = 600
>>
>> #IQSIM Orphan SMPP Connector
>> group=smsc
>> smsc=smpp
>> smsc-id=internal
>> interface-version=34
>> host=##########
>> receive-port=2775
>> system-id=sms2
>> smsc-username=########
>> smsc-password=########
>> system-type=default
>> #max-pending-submits=70
>> #validityperiod=20160
>>
>> #IQSIM Sending SMPP Connector
>> group=smsc
>> smsc=smpp
>> smsc-id=internal
>> interface-version=34
>> host=############
>> port=2775
>> system-id=sms1
>> smsc-username=#######
>> smsc-password=#######
>> system-type=default
>> max-pending-submits = 410
>> #validityperiod=20160
>> transceiver-mode=1
>> denied-smsc-id=#######
>>
>> group = smsbox
>> bearerbox-host = 127.0.0.1
>> sendsms-port = 13013
>> global-sender = 13013
>> smsbox-id=my_smsbox
>>
>> #Routing for standard Kannel messages i.e. no smsc_id
>> group=smsbox-route
>> smsbox-id=sqlbox
>> smsc-id=internal
>>
>>
>> #--------------------------------------------------------------------
>> #SEND-SMS USERS
>> #--------------------------------------------------------------------
>>
>> group = sendsms-user
>> default-smsc = default
>> username = playsms1
>> password = playsms1
>> concatenation = true
>> max-messages = 3
>>
>> group = mysql-connection
>> id = mydlr
>> host = 154.72.8.195
>> username = root
>> password = at0m1c55
>> database = kannel
>> #max-connections = 20
>>
>> group = dlr-db
>> id = mydlr
>>
>> On 13 Jan 2016, at 13:20, Otandeka Simon Peter <sotand...@gmail.com>
>> wrote:
>>
>>
>> Hi,
>>
>> I have used a couple of distros (centos, freebsd ubuntu, wheezy,
>> openSuSE) and they all work well without issues.  It's how you configure
>> your system (resources) that matters a lot i.e. apache, kannel, memory etc.
>>
>> You may want to share some bearerbox logs and config so that we can
>> advise on the retransmission bit.
>>
>> @sotandeka
>>
>> On Wed, Jan 13, 2016 at 2:03 PM, Grant Saicom <grant.sai...@gmail.com>
>> wrote:
>>
>>> Hi Everyone
>>>
>>> I am looking for an opinion (good natured discussion please) on which
>>> distribution is best to run a lean purpose built kannel system for smpp sms
>>> delivery and replies using sqlbox.
>>>
>>> DB is mysql and the DB is running on a separate VM on the same
>>> hypervisor with SSD drives.
>>>
>>> We are currently running debian wheezy 7.8 installed using net install
>>> so it was as minimal as possible. We are seeing quite a bit of tcp
>>> retransmission going on between the system and the smsc.
>>>
>>> Can anyone advise from their own experience?
>>>
>>> Kind regards
>>> Grant
>>>
>>
>>
>>
>>
>

Reply via email to