The problem is not reproducible in the clean environment :( I have re-opened https://issues.apache.org/jira/browse/OPENMEETINGS-2192 But we need to figure out reproducible steps for this :(
On Thu, 26 Mar 2020 at 07:43, Rodolfo Gonzalez <rodolfo.gonza...@gmail.com> wrote: > Hi Maxim, the problem is, in the server output I can not find exceptions > other than those refused logins. In the client (PHP) side, I only get that > the object which is returned by the login API call is null (so the call to > get the room hash crashes, as the method is called from a null). > > I mean, I find other kind of exceptions, concerning to video stream and > other things. But the problem seems to be that the login is being rejected. > Another user posted a similar (maybe the same) problem just a moment ago. > > Regards. > > El mié., 25 mar. 2020 a las 18:21, Maxim Solodovnik (<solomax...@gmail.com>) > escribió: > >> Hello Rodolfo, >> >> Debug logging you quoted provide zero info :( >> There are no limits for logins in the system >> >> Could you share error stack trace? >> >> I might write test foe this >> Am I right thinking the use case is: 100 users are trying to get room >> hash in parallel? >> >> >> On Thu, 26 Mar 2020 at 00:58, Rodolfo Gonzalez < >> rodolfo.gonza...@gmail.com> wrote: >> >>> Hi, >>> >>> currently I run an OM 4.0.10 server with a pretty default configuration, >>> using >>> >>> # java -version >>> openjdk version "1.8.0_242" >>> OpenJDK Runtime Environment (build >>> 1.8.0_242-8u242-b08-0ubuntu3~16.04-b08) >>> OpenJDK 64-Bit Server VM (build 25.242-b08, mixed mode) >>> >>> on Ubuntu 16.04 (default values for file handles limits). >>> >>> I integrated OM with a web app made in PHP with the Yii framework, using >>> the SOAP API. I use this library: >>> >>> https://github.com/hakeem23/open-meetings-soap >>> >>> The server is on AWS, being a c4d.4xlarge instance, with htop reporting >>> 16 cores. CPU usage is reported by AWS to be around 50%, with a system load >>> maxing up at the 5-7 range reported by htop. The thing is that as many as >>> 100 users try to list rooms or connect to the server at a given time. >>> >>> I'm having several problems with the service, the main one being that >>> the users sometimes can not get the room hashes or enter a room. Given the >>> flow of the process, my best guess is that the library can not login to the >>> OM server. The errors in the output are like these, I think (I'm not sure >>> if these are the cause of the problem): >>> >>> DEBUG 03-25 11:44:24.577 o.a.o.d.u.AuthLevelUtil:42 [0.0-5453-exec-7] - >>> Level Admin :: [DENIED] >>> DEBUG 03-25 11:44:24.632 o.a.o.d.u.AuthLevelUtil:42 [0.0-5453-exec-7] - >>> Level Admin :: [DENIED] >>> DEBUG 03-25 11:44:24.633 o.a.o.d.u.AuthLevelUtil:42 [0.0-5453-exec-7] - >>> Level GroupAdmin :: [DENIED] >>> DEBUG 03-25 11:44:24.642 o.a.o.d.u.AuthLevelUtil:42 [0.0-5453-exec-7] - >>> Level Admin :: [DENIED] >>> DEBUG 03-25 11:44:24.642 o.a.o.d.u.AuthLevelUtil:42 [0.0-5453-exec-7] - >>> Level Admin :: [DENIED] >>> >>> and so on. But I'm unsure if this is the error which causes a cascade >>> exceptions which render the web app unable to get the room hashes and so >>> (those process relay on being loged in, anyway). The thing is, the SOAP >>> library currently does the login process from the app, using and user >>> "admin" with SOAP and ADMIN level credentials. And the problem is >>> intermitent. >>> >>> Do you consider that my guess is right? Is there a limit on "logins" at >>> any given time? If so, can it be increased? Does the version of java affect? >>> >>> Any other ideas are welcome. >>> >>> Regards, >>> Rodolfo. >>> >> >> >> -- >> WBR >> Maxim aka solomax >> > -- WBR Maxim aka solomax
