| CPU leak in nowsms |  Search
|
|
NowSMS Support Forums ⬆ NowSMS Support - SMS Issues ⬆ Archive through September 15, 2003 ⬆ |
◄ ► |
| Author | Message | |||
| lars mservice |
ExtraHTTPargs in nowsms is a fantastic feature. But when traffic is high, like 5 messages in a row, then smsgws.exe uses 100% cpu! and i cannot reach the server via pc anywhere or other remote stuff. Only way to stop it, is to restart the server. My smsdebug.log says absolutely nothing Anyone out there been seeing the same ? | |||
| Bryce Norwood - NowSMS Support |
Lars, As Kent mentioned, I think we would need to see your SMSDEBUG.LOG to see if there is anything that we could notice which is unusual. We'd want to see the URLs being used for the messages that were posted ... try to recreate the problem ourselves with configuration settings similar to yours. I suspect that one of the messages being submitted is the one triggering a problem ... and somehow the gateway is in a hard loop processing the extra HTTP parameters related to the request. -bn My guess is that in this case there is something about the parsing of the ExtraHTTPArgs parameter in SMSGW.INI that is causing the parser to loop when trying to process a message. Make sure there is no white space in the parameter setting. | |||
| lars mservice |
Hi Bryce. I consider myself to be quite experienced in reading the smsdebug.log  I read it every day just to check if everything is allright. What happends when smsgws.exe uses 100% cpu is that it continues running perfect. It just takes 2 minutes for the messages to be processed and sent out through nowsms. I have checked and checked the logfiles, but cannot find anything wrong. It only booms the cpu if i post for example 5 messages in a for next loop. That means 5 posts fast to nowsms. If i only post 3 messages, nothing happends. Nowsms runs smoothly (75% cpu in about 30 seconds). I use one file to post to nowsms. It is the one file that i have used for a long time now, and i dont suspect that to be wrong. I will check everything again, but i really dont think i will find anything wrong. Try setting up extrahttpargs on your testserver, and the try posting 5 or 10 messages to nowsms through a for next loop. See what happends on your server. Remember that i post from another server than the one running nowsms. Here is some stuff from my smsgw.ini: ExtraHttpArgList=mediacode,appnr,sessionid TemplateText=/push?to=@@PhoneNumber@@&text=@@Text@@&smsc=telia&appnr=@@appnr@@&m ediacode=@@mediacode@@ TemplateBinary=/push?to=@@PhoneNumber@@&type=extended&text=@@Data@@&udh=@@UDH@@& smsc=telia&appnr=@@appnr@@&mediacode=@@mediacode@@ | |||
| Bryce Norwood - NowSMS Support |
Lars, For the CPU to get stuck at 100%, some process would have to get stuck in a loop. I could only figure that a thread that was processing a message gets stuck in a loop parsing those extra variables. That's where I'm stumped in your scenario. I would have to figure that the system is getting into a loop involving the processing of one of your messages. Otherwise, I don't know what it could be stuck processing. That's why I figure we could at least track through a log to verify that all of the messages do actually get routed out. Something has got to be stuck for the processor to be at 100% ... I just can't see how anything could be affected that wasn't actively processing a message. -bn | |||
