Re: Smsbox sometimes doesn't servicing request

Try with real SMSC connection simulation, HTTP will be the easiest. And my case I used Kannel HTTP SMSC and the bug still exists.


On Dec 13, 2012, at 8:27 PM, Rene Kluwen wrote:

> I used version svn head in combination with fakesmsc.
>  
> From: Willy Mularto [mailto:[email protected]] 
> Sent: donderdag 13 december 2012 2:07
> To: Rene Kluwen
> Cc: '[email protected]'
> Subject: Re: Smsbox sometimes doesn't servicing request
>  
> Which Kannel version did you use? Have you tried HTTP SMSC?
>  
>  
>  
> On Dec 12, 2012, at 10:11 PM, Rene Kluwen wrote:
> 
> 
> I tested this issue with:
>  
> test/fakesmsc -r 13006 -i 0.00001 -m 50000 "123 999 text itn2"
>  
> For testing purposes, I set up Kannel the following way:
>  
> <http script> -- <smsbox> -- <sqlbox1> -- <sqlbox2> -- <bearerbox> -- <fakesmsc>
>  
> After the test, the following is true:
>  
> Messages sent to sqlbox2: 100,000
> Messages sent to sqlbox1: 100,000
> Messages received by php script: 40,775
> Smsbox http errors: (Could not fetch content): 46
> PHP errors (too many concurrent connections to database): 9,179
>  
> This means that I cannot reproduce this so-called sqlbox bug.
> Sqlbox perfectly forwards all messages, considering both sqlbox1 and sqlbox2 have the proper message count.
> The reason why some messages _seem_ to be lost is because of http errors in smsbox AND counting the messages in the php script.
>  
> Also, considering we’ve got a reply from smsbox to all messages (message count is 100,000), it shows there is no bug in either sqlbox or smsbox.
>  
> If you add all the messages above together (40,775 + 46 + 9,179) then the total is 50,000. So no messages disappear suddenly.
>  
> == Rene
>  
>  
> From: Willy Mularto [mailto:[email protected]] 
> Sent: woensdag 12 december 2012 1:37
> To: Rene Kluwen
> Subject: Re: Smsbox sometimes doesn't servicing request
>  
> I enabled kannel internal http smsc interface. That's the easiest way to test. First I test the MT, Kannel 100% can do it. But MO sometimes more than 50% failed. What I mean failed is, bearerbox 100% received the MO (proven by checking the log and compare it the number of MO test), but SMSBox didn't receive all MO. It means the MO lost between Bearerbox=>SMSBox and as we know SQLBox listens to Bearerbox and (should) forward to SMSBox.
>  
>  
>  
> On Dec 11, 2012, at 9:13 PM, Rene Kluwen wrote:
> 
> 
> 
> But my question was: What is your test-setup?
> I would like to test the same way as you did. So I can be sure that –when I make a patch- that it works.
>  
> == Rene
>  
> From: Willy Mularto [mailto:[email protected]] 
> Sent: dinsdag 11 december 2012 1:49
> To: Rene Kluwen
> Subject: Re: Smsbox sometimes doesn't servicing request
>  
> Great, please let me know your result. 
>  
>  
>  
> On Dec 10, 2012, at 6:52 PM, Rene Kluwen wrote:
> 
> 
> 
> 
> No, it’s just that until now I start testing this.
>  
> == Rene
>  
>  
> From: Willy Mularto [mailto:[email protected]] 
> Sent: maandag 10 december 2012 1:36
> To: Rene Kluwen
> Subject: Re: Smsbox sometimes doesn't servicing request
>  
> Is this a delayed message?
>  
>  
>  
> On Dec 10, 2012, at 12:05 AM, Rene Kluwen wrote:
> 
> 
> 
> 
> 
> Maybe this error is due to smsbox.
> Could you give me your test setup, so I can run some tests?
>  
> == Rene
>  
> From: Willy Mularto [mailto:[email protected]] 
> Sent: dinsdag 23 oktober 2012 8:25
> To: Rene Kluwen
> Cc: Alexander Malysh; kannel [email protected]; Stipe Tolj
> Subject: Re: Smsbox sometimes doesn't servicing request
>  
> Hi list,
> After doing another MO scenario which only run Bearerbox and SMSbox. I see no bug, tested with 411402 MO via HTTP Generic SMSGW all received by Bearerbox and also SMSBox well, check other end point get-url script and can confirm it received all MO request. Then try to put SQLBox in chain and the bug start to happen. So I think developer needs to rewrite SQLBox to accommodate the current Bearerbox and SMSBox architecture. Hope this bug solved soon. I need confirmation from other users also. Thanks.
>  
>  
>  
> On Sep 26, 2012, at 6:02 PM, Willy Mularto wrote:
> 
> 
> 
> 
> 
> 
> Hi Rene,
> This test involved only 1 SMSC and 1 SMSBox :)
>  
>  
> On Sep 26, 2012, at 5:48 PM, Rene Kluwen wrote:
> 
> 
> 
> 
> 
> 
> I have a feeling where to look for this.
> Willy, in the test below: Are we talking about 1 smsc and 1 smsbox? Or multiple?
>  
> == Rene
>  
>  
> From: Willy Mularto [mailto:[email protected]] 
> Sent: Wednesday, 26 September, 2012 03:17
> To: Rene Kluwen
> Cc: 'Alexander Malysh'; 'kannel [email protected]'
> Subject: Re: Smsbox sometimes doesn't servicing request
>  
> MO appears in the sent_sms correctly. Test with 10K MO all in bearerbox-access.log and also all in sent_sms. But some and I prefer say many actually, on repeated of 10K tests only 4K-5K  go through SMSBox. 
>  
>  
> On Sep 26, 2012, at 3:17 AM, Rene Kluwen wrote:
> 
> 
> 
> 
> 
> 
> 
> So the question here is: Do the MO’s appear in sent_sms, but didn’t go through to smsbox?
>  
> == Rene
>  
> From: [email protected] [mailto:[email protected]] On Behalf Of Alexander Malysh
> Sent: Tuesday, 25 September, 2012 15:33
> To: Willy Mularto
> Cc: kannel [email protected]
> Subject: Re: Smsbox sometimes doesn't servicing request
>  
> Hi,
>  
> do I read this correct, the way smsbox talk to sqlbox? or do you mean bearerbox talk to sqlbox?
>  
> Alex
>  
> On 21.09.2012, at 03:34, Willy Mularto <[email protected]> wrote:
> 
> 
> 
> 
> 
> 
> 
> 
> Hi Alex,
> I did some researches on this matter and can confirm that this issue started right after meta-data was implemented (r4682). And It is more the way SMSBox talk to SQLBox and vice versa. Today I am going to compile current SQLBox SVN against r4681 and let's see the result.
>  
>  
>  
> On Sep 20, 2012, at 11:23 PM, Alexander Malysh wrote:
> 
> 
> 
> 
> 
> 
> 
> 
> Hi,
>  
> is this only sqlbox bug or smsbox as well?
>  
> Thanks,
> Alex
>  
> On 20.09.2012, at 04:42, Willy Mularto <[email protected]> wrote:
> 
> 
> 
> 
> 
> 
> 
> 
> Yes, all use the latest SVN. I can confirm this is a bug and an old bug actually (since meta-data implemented). Already posted to redmine and contacted developers. Hope they can found the solution.
>  
>  
>  
> On Sep 19, 2012, at 10:47 PM, spameden wrote:
> 
> 
> 
> 
> 
> 
> 
> 
> Hey Willy!
> 
> Did you use latest version from SVN?
> 
> 2012/9/19 Rene Kluwen <[email protected]>
> You seem to have crossed an old sqlbox bug.
> I remember, under high load, sqlbox misses some MO messages.
> 
> As of now, for all I know, this hasn't been solved yet.
> 
> == Rene
> 
> -----Original Message-----
> From: [email protected] [mailto:[email protected]] On
> Behalf Of Willy Mularto
> Sent: Wednesday, 19 September, 2012 12:35
> To: kannel [email protected]
> Subject: Smsbox sometimes doesn't servicing request
> 
> Guys,
> I have a weird situation. smsbox sometimes doesn't servicing request (not
> forward MO to get-url). Here is my configuration:
> 
> group = core
> admin-port = 7960
> smsbox-port = 7961
> admin-password = kannel
> status-password = kannel
> box-allow-ip = "127.0.0.1;192.168.8."
> store-type = spool
> store-location = "/home/kannel/var/queue"
> sms-incoming-queue-limit = -1
> sms-outgoing-queue-limit = -1
> store-dump-freq = 5
> dlr-storage = mysql
> sms-resend-freq = 60
> sms-resend-retry = -1
> smsbox-max-pending = 100
> log-file = "/home/kannel/var/log/bearerbox.log"
> access-log = "/home/kannel/var/log/bearerbox-access.log"
> log-level = 0
> 
> group = mysql-connection
> id = mydlr
> host = 192.168.8.3
> username = "kannel"
> password = "password"
> database = traffics
> 
> group = dlr-db
> id = mydlr
> table = partner_dlr
> field-smsc = smsc
> field-timestamp = ts
> field-destination = destination
> field-source = source
> field-service = service
> field-url = url
> field-mask = mask
> field-status = status
> field-boxc-id = boxc
> 
> group = smsc
> smsc = http
> system-type = kannel
> smsc-id = kannel
> smsc-username = kanneluser
> smsc-password = kannelpass
> port = 7278
> send-url = "http://192.168.8.2/test.php"
> connect-allow-ip = "127.0.0.1;192.168.8.2"
> allowed-smsc-id = "kannel"
> log-file = "/home/kannel/var/log/partner.log"
> log-level = 0
> 
> group = smsbox-route
> smsbox-id = "smsbox"
> smsc-id = "kannel"
> 
> group = sendsms-user
> username = kannel
> password = kannel
> max-messages = 10
> concatenation = true
> 
> group = sms-service
> keyword = default
> get-url =
> "http://handler.kannel.gw:4684/?sender=%p&receiver=%P&sms=%a&telco=%i&msgid=
> %I&fid=%F"
> catch-all = true
> omit-empty = true
> max-messages = 0
> 
> group = smsbox
> smsbox-id = smsbox
> sendsms-port = 7962
> bearerbox-port = 7963
> http-request-retry = 100
> http-queue-delay = 10
> max-pending-requests = 8192
> mo-recode = 1
> log-file = "/home/kannel/var/log/smsbox_1.log"
> access-log = "/home/kannel/var/log/smsbox-access_1.log"
> log-level = 1
> 
> group = mysql-connection
> id = smsgw-db
> host = 192.168.8.3
> port = 3306
> username = "kannel"
> password = "password"
> database = "traffics"
> max-connections = 10
> 
> group = sqlbox
> id = smsgw-db
> bearerbox-port = 7961
> smsbox-id = smsbox
> smsbox-port = 7963
> sql-log-table = partner_sent
> sql-insert-table = partner_queue
> log-file = "/home/kannel/var/log/sqlbox_1_pull.log"
> log-level = 0
> 
> Willy Mularto
> F300HD+MR18DE (NLC1725)
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
>  
>  
> Willy Mularto
> F300HD+MR18DE (NLC1725)
>  
>  
>  
>  
>  
>  
> 
> 
> 
> 
> 
> 
> 
> 
>  
>  
>  
> Willy Mularto
> F300HD+MR18DE (NLC1725)
>  
>  
>  
>  
>  
>  
> 
> 
> 
> 
> 
> 
> 
> 
>  
>  
> Willy Mularto
> F300HD+MR18DE (NLC1725)
>  
>  
>  
>  
>  
>  
> 
> 
> 
> 
> 
> 
> 
>  
>  
> Willy Mularto
> F300HD+MR18DE (NLC1725)
>  
>  
>  
>  
>  
>  
> 
> 
> 
> 
> 
> 
>  
>  
> Willy Mularto
> F300HD+MR18DE (NLC1725)
>  
>  
>  
>  
>  
>  
> 
> 
> 
> 
> 
> 
>  
>  
> Willy Mularto
> F300HD+MR18DE (NLC1725)
>  
>  
>  
>  
>  
>  
> 
> 
> 
> 
> 
>  
>  
> Willy Mularto
> F300HD+MR18DE (NLC1725)
>  
>  
>  
>  
>  
>  
> 
> 
> 
> 
>  
>  
> Willy Mularto
> F300HD+MR18DE (NLC1725)
>  
>  
>  
>  
>  
>  
> 
> 
> 
>  
>  
> Willy Mularto
> F300HD+MR18DE (NLC1725)
>  
>  
>  
>  
>  
>  
> 
> 
>  

Willy Mularto
F300HD+MR18DE (NLC1725)