Forum Replies Created
Is there a different email I should send those logs to?
I just sent php error logs, slow logs and database layer CPU usage for the last 7 days via WeTransfer to support@wpdownloadmanager.com.
For reference, Flywheel’s database layers already have an extremely high CPU spec, so seeing the usage spiking to 100% sustained usage is a bit surprising. They suggest working with WPDM developers to reduce the database usage as a recommended first step as problems only occur when the usage is pinned at 100%, which you can see happens somewhat regularly.
I haven’t changed the login credentials. Please try again. Reminder that the login URL isn’t in the original request but in the response following. I can post it again below to keep it together.
For further information, our host (Flywheel) has been able to determine the following during our down periods:
We do see that the site is causing a sustained level of 100% CPU usage on the Database layer. When we dig into this further we can see that around the same time as these spikes happen, the plugin download-manager is throwing thousands and thousands of slow errors. Slow errors are processes that take longer than 5 seconds to complete. These errors can hog site resources and are a common suspect for performance and downtime issues. You may want to reach out to the developers with these errors as they may have more specific troubleshooting to mitigate these or identify any potential conflicts.
To give you an idea it has thrown 8500 of these errors in the last 24 hours. Another plugin, wpdm-pdf-viewer has thrown about 6400 and may also need further investigation. Do you have load testing abilities to try to simulate the experience we are having?
If you happen to be online now, it is happening right now.
Yes I see it is working currently, but as mentioned, when we see an increase in site traffic it stops working. When you are getting in the site to troubleshoot is when our clients are offline and not doing business so it is hard to replicate the issue.
Also, I totally understand the need to make test packages to test, but when you are finished could you change them to draft? Our site is live to the public and I don’t want our test packages to be visible to users.
What further steps can we take to investigate since we are unable to get in touch with your team in the moment that we are experiencing issues?
Since it appears the initial fatal error has been resolved, I’m going to remove the admin access to the site. We are currently using a different method as a workaround to the issues we’ve been having. It will be difficult for you to thoroughly test what we have been experiencing with that workaround in place and not being able to see live during those live traffic times what is occurring. If I’m able to recreate the issue and tie it to slow logs that say what might be happening, I’ll reach back out. But right now I think we are maybe at an stopping spot. Thanks for your help.
Yes, the whole site is SSO protected and cannot be viewed without using the login credentials provided in the previous private post. Please let me know if that works.
Yes, we are still having issues. PDFs have not been loading for over an hour. In chatting our host, they see a massive amount of slow errors coming from download-manager and wpdm-pdf-viewer.
We have installed the update, but are still receiving errors with PDFs loading and giving an error that says:
An error occurred while loading the PDF.
Does seem to be intermittent still. Have been unable to confirm if the fatal error is still occurring.
I do see v3.1.0 out now, so I’ll get it updated. Is that the version I should be expecting to see this change?
Thank you, Nayeem. Do you have an estimate on when that update might be available?
I am unable to turn this on for a few reasons…for one the site is live to the public and turning on debug mode isn’t a good experience on the frontend. However, I have downloaded the logs from the past 2 days. Is there a way that I can share that file with you? I don’t have access to Dropbox, etc. so I’m hoping you have a file share system that I can send that to you.
We are having issues again currently and we continue to get the following error from your plugin:
NOTICE: PHP message: PHP Fatal error: Uncaught TypeError: method_exists(): Argument #1 ($object_or_class) must be of type object|string, null given in /www/wp-content/plugins/wpdm-pdf-viewer/tpls/viewer.php:68
As mentioned previously, this is an intermittent issue. When it appears, our host is able to pull slow logs and see fatal errors occurring with your plugin. I can reach back out the moment it occurs again if that is helpful. Are you able to pull logs and see these errors yourself?
Is there a development team that we can elevate this to? Download Manager is the core functionality of our site and we have to get this resolved or else come up with a plan to replace this plugin with another solution which will take time that we don’t have while things are still not working intermittently.
I appreciate you getting in to look around, I’d just ask if you don’t mind keeping any test content published as a draft following testing, especially during our business hours. We want to avoid having test content out for people to access.
One more update before I leave for the day as we experienced downtime again today. Flywheel was able to tell us the following at the time of the incident:
For the PDF viewer, it looks like that plugin is generating a Fatal Error:
PHP Fatal error: Uncaught TypeError: method_exists(): Argument #1 ($object_or_class) must be of type object|string, null given in /www/wp-content/plugins/wpdm-pdf-viewer/tpls/viewer.php:68
Fatal errors like this will break functionality and can also cause quite a slowdown as code isn’t executing properly. We recommend working with the plugin authors to get that patched up at your earliest convenience.
In addition to the Fatal Error, this plugin is also generating a large number of slow log entries, as is the download-manager plugin. Its possible these two issues are connected. It is generating a huge number more slow log entries than it was yesterday (about 7,000 in the last 24 hours alone).
Okay thank you. Let me know if you have any trouble accessing.