I'm not necessarily expecting an answer here because it makes little sense. But I wanted to report in case someone else does run across this and has some insight. I'm running a simple shopfloor time punch app on a Chromebox in current version of Chrome browser. It does not run through VVcall, just a direct call to server program.
It worked fine for several weeks until last Thursday. I had an error in another Valence app which I didn't handle and it essentially killed all the CGI jobs avail so Valence wouldn't run at all. I fixed the error next morning and thought all was well. Then Supervisors reported the punch app was running very slow, taking like 20-30 seconds to punch someone in.
After I verified it wasn't a network issue, I went over to look at problem. Using debugger, I could see that my ajax call was hanging as "pending" for 20 seconds before it even made the POST to server. On web, I've seen many developers complain about this in Chrome for years. But in their case, it was repeatable. In my case, it was only this one Chromebox. All other Chrome browsers, running on Windows computers, could use this app just fine. This Chromebox was connected to our local LAN with no internet access.
I tested with other Chromebox I have, no issues. I discovered if I did a hard power off, the app would work correctly one time before the "pending" would come back. I reset the failing Chromebox to factory default and recreated the account, now the error appears to be gone.
Only thing that makes sense is Chrome os somehow remembered it did not get a response when the CGI jobs were down on i5. Then it held this request as "pending" until it tried to resolve some things, maybe even trying to get to internet to report to Google. I've now turned off all the stuff I could find in Advanced Settings to prevent that behavior in future.
If anyone has any insight on this, please let me know. It's just a little scary because my plan is to run 50 of these devices on shop floor making XHR requests, this would be disaster. Thanks!
It worked fine for several weeks until last Thursday. I had an error in another Valence app which I didn't handle and it essentially killed all the CGI jobs avail so Valence wouldn't run at all. I fixed the error next morning and thought all was well. Then Supervisors reported the punch app was running very slow, taking like 20-30 seconds to punch someone in.
After I verified it wasn't a network issue, I went over to look at problem. Using debugger, I could see that my ajax call was hanging as "pending" for 20 seconds before it even made the POST to server. On web, I've seen many developers complain about this in Chrome for years. But in their case, it was repeatable. In my case, it was only this one Chromebox. All other Chrome browsers, running on Windows computers, could use this app just fine. This Chromebox was connected to our local LAN with no internet access.
I tested with other Chromebox I have, no issues. I discovered if I did a hard power off, the app would work correctly one time before the "pending" would come back. I reset the failing Chromebox to factory default and recreated the account, now the error appears to be gone.
Only thing that makes sense is Chrome os somehow remembered it did not get a response when the CGI jobs were down on i5. Then it held this request as "pending" until it tried to resolve some things, maybe even trying to get to internet to report to Google. I've now turned off all the stuff I could find in Advanced Settings to prevent that behavior in future.
If anyone has any insight on this, please let me know. It's just a little scary because my plan is to run 50 of these devices on shop floor making XHR requests, this would be disaster. Thanks!
Comment