Briggs Softworks

https://gyazo.com/8436c3838720e7e7299888d400a72558

Both address 007B8826 showing 6300200A as well as 004C89D0 showing 00001DFF showed up several times upon checking the client the day after we patched it, could you please explain what is causing this and if it's an issue at all, if so how to prevent/fix it? Thank you for for your time

And now another one is popping up constantly, making me unable to use the server interface. https://gyazo.com/862e3e72e73f5f61b8f26d5db8b9fae5 004C951F : 8EF80008 this time.

Is that the Pro or Gold edition? Is there anything recorded in your Error Logs?

I noticed from this thread that were having similar issues in 5.16: http://www.briggsoft.com/forums/viewtop ... =10&t=2616

So did that issue go away in 5.20 and then reappear in 5.21? Also I can tell from the address length in the previous thread that you were running the 64-bit edition but now are using the 32-bit edition. Is there a reason for that?

According to my compiler's map file on the 32-bit Gold 5.21 version, address 007B8826 occurs at a point when the code is adding a string to a list. The only thing that should trigger an error at that point is if you were out of memory or out of system resources. How many logins did you have at around that time? Is your site protected by cloudflare? A denial of service attack could also exhaust all of the server's memory.

We are running PM 5.21 gold 64 bit, the issue appeared previously but not to this extent, now there is 20+ of these errors daily and various new ones are included, the server has crashed since getting around 60 of these, error logs show nothing interesting surrounding the crash time (Always the case) however I do see a ton of these appearing throughout the day:
2018-04-02 03:10:00|Callback event error: Read timed out.
2018-04-02 03:10:01|Callback events resumed
2018-04-02 04:16:08|Callback event error: Read timed out.
2018-04-02 04:16:09|Callback events resumed
2018-04-02 05:09:52|Callback event error: Read timed out.
2018-04-02 05:09:55|Callback events resumed
2018-04-02 06:12:37|Callback event error: Read timed out.
2018-04-02 06:12:38|Callback events resumed
2018-04-02 14:22:46|Callback event error: Read timed out.
2018-04-02 14:22:50|Callback events resumed
2018-04-02 15:22:11|Callback event error: Read timed out.
2018-04-02 15:22:12|Callback events resumed
2018-04-02 16:15:25|Callback event error: Read timed out.
2018-04-02 16:15:26|Callback events resumed

As well as quite a few Failed API attempts from xyz IP addresses, these are the only error logs showing up consistantly

needsupport22 wrote:We are running PM 5.21 gold 64 bit

Those are 32-bit addresses in your screenshot. Check the title bar on the server console. The 64-bit edition will show "64-bit" on the end. Otherwise it's the 32-bit edition.

In any case, monitor "PMServer.exe" on the Processes tab in the Windows Task Manager and keep an eye on the CPU, Memory, and Thread columns. Or setup a Perfmon log to record those values every second to a log file. Instructions here:
http://www.briggsoft.com/forums/viewtop ... 8674#p8674

I don't get how it could have possibly patched from 64 bit into 32 bit, because this happened then according to your info.

needsupport22 wrote:I don't get how it could have possibly patched from 64 bit into 32 bit, because this happened then according to your info.

The patch extracts the full setup program and runs it. There you'll see the option to do a 32-bit or 64-bit install. It should default to whatever you ran previously but if it didn't and you weren't paying attention it could install the other one. You can just run that setup file (not the patch) again and switch it back to 64-bit. Search your drive for pg521setup.exe

Thank you Kent I'll relay to my dev, that hopefully does the trick