Connection error when scrolling all articles and fresh section


#1

Hello,just recently stumbled across some odd behaviour while scrolling through the all articles and fresh

section with view all articles set, after the section updates some 5-6 times it comes to a halt and feed locks up

and wont refresh then times out some 3 minutes while giving the yellow triangle with “communication error”, it

locks up at same place always in browser,i thought it might could be some query limit on apache or db or so,

but it only applies to above mention sections, my main feed

category i can scroll endlessly without it ocurring,also this only applies to browser, ttrss android app i cant

reproduce while scrolling past the articles where it locks on browser. While i can just scroll through the

“folder/category” i placed most of my feeds without problems im hesitant reporting this as it doesnt cripple the

functionality to scroll through all feeds,also tried different browser with same results. Im using systemd

daemon to update and also using syslog and

cant find any errors there,also tried httpd error log couldnt find any errors related, Just thought i should report

though also again want to throw thanks for this awsome software its made my day alot better regarding to

news :slight_smile: also apologise in advance if post is off in some terms, i dont frequent forums often…

System: centos 7, php 7.2.10, mariadb 5.5.60, Ttrss v18.8 (df0115f)


#2

Sounds like MariaDB is getting stuck on some query or PHP is having trouble.

Login to TT-RSS as the admin and in Preferences view the log to see if any PHP errors are logged. If not, then in config.php change the log destination to an empty string and re-create the issue. Then check on the server wherever PHP error logs are stored (depends on what you have set in your system’s PHP config files).

You can also enable the slow query log in MariaDB (search the Internet for steps).

One thing to also try, SSH into your server and run top, then re-create the issue and observe if there are CPU or memory spikes for any of the processes. It might help narrow down the issue.


#3

Thanks, i will report my findings next few days


#4

Thats weird, i changed log destination from syslog to sql and now i cant reproduce. It would seem possibel if it

was the other way around that it would time out due to queries or so but for syslog id think maybe not. I’ll tinker

some more with it and report back,so with sql logging no problemo :slight_smile:


#5

By default TT-RSS logs errors to the database, but if the error is with the actual database it can sometimes fail in logging the error as well. fox has made some improvements to this recently, but if the database itself is the problem (e.g. too many connections, etc.) it obviously isn’t going to get logged.

I suggest running with PHP logging to a file for a few days, if the timeout occurs again then check the error logs to see if you can find where the problem is and report back.