Feeds Don't Get Automatically Refreshed

Hello.

I’ve recently updated my TT-RSS installation. However, before the update, I was having an issue with feed refreshing that feeds wouldn’t get automatically refreshed as they used to. I thought an update would solve the issue but it’s still ongoing.

The issue, as far as I know, appeared right out of the blue, perhaps an update on my Linux box(CentOS 7 with Apache) caused the issue but I haven’t touched my TT-RSS installation myself.

I’m making use of the systemd to initiate the feed refreshes. The file is located at /etc/systemd/system/.

Currently, the service for TT-RSS is active and running. Here is the current output of systemctl status ttrss_backend:

● ttrss_backend.service - ttrss_backend
   Loaded: loaded (/etc/systemd/system/ttrss_backend.service; enabled; vendor preset: disabled)
   Active: active (running) since Sun 2020-01-19 09:12:55 CET; 1 weeks 3 days ago
 Main PID: 24115 (php)
   CGroup: /system.slice/ttrss_backend.service
           └─24115 php /var/www/rss.cansurmeli.com/public_html/update_daemon2.php

Jan 29 09:14:20 vps update_daemon2.php[24115]: [08:14:20/20950] Removing old archived feeds...
Jan 29 09:14:20 vps update_daemon2.php[24115]: [08:14:20/20950] Purged 1 orphaned posts.
Jan 29 09:14:20 vps update_daemon2.php[24115]: [08:14:20/20950] Removed 0 (feeds) 0 (cats) orphaned counter cache entries.
Jan 29 09:14:21 vps update_daemon2.php[24115]: [08:14:21/24115] [reap_children] child 20948 reaped.
Jan 29 09:14:21 vps update_daemon2.php[24115]: [08:14:21/24115] [SIGCHLD] jobs left: 1
Jan 29 09:14:22 vps update_daemon2.php[24115]: [08:14:22/20951] Scheduled 0 feeds to update...
Jan 29 09:14:22 vps update_daemon2.php[24115]: [08:14:22/20951] Sending digests, batch of max 15 users, headline limit = 1000
Jan 29 09:14:22 vps update_daemon2.php[24115]: [08:14:22/20951] All done.
Jan 29 09:14:23 vps update_daemon2.php[24115]: [08:14:23/24115] [reap_children] child 20949 reaped.
Jan 29 09:14:23 vps update_daemon2.php[24115]: [08:14:23/24115] [SIGCHLD] jobs left: 0

The only way I can initiate an update is getting into VPS, going into the appropiate directory and calling php update.php --force-update. Then, after a while, I get the updated feeds on the web side of TT-RSS.

As far as I can see, there are no errors or other inconviniences. What could be the issue here?

Thanks.

  1. read the sticky for the support subsection and fill in everything that is missing.
  2. the usual cause for this is system time jumping around. alternatively, this could be something related to your database (i have a vague recollection of some particularly broken mariadb versions breaking tt-rss updates - search the forum).
  3. if you can, switch to the docker setup.

I went for a cron job that runs every 15 minutes instead of the daemon, because you never know if an software or OS update brakes anything.
Works reliable for years…

this won’t really change anything, you’re just running exactly same updating code in a slightly different way.

Sorry about the missing information, my bad. Here are they:

  • CentOS Linux release 7.7.1908 (Core) on a VPS (DigitalOcean Droplet),
  • 5.5.64-MariaDB MariaDB Server,
  • PHP 7.2.27 (cli) (built: Jan 22 2020 09:31:55) ( NTS ),
  • php-opcache was not installed, now it is.

I realise that in the requirements it’s mentioned that MySQL isn’t directly supported but at the time I had a MariaDB setup so I went on with it.

I knew that my system time was correct. Although, I checked it again and was normal.

However, I did another git pull. I’m on the Mon Jan 27 15:45:04 2020 +0300 commit now. Now refreshing works.

I just wanted to publish the details for other users if it could help them in the future.

Ok, another update. My previous observation that my TT-RSS instance got back to regular updating of the feeds on it’s own was wrong. It was just that some feeds were being updated and it led me to think that the issue got resolved on it’s own but I quickly realised it afterwards.

I just didn’t want to write something else before having a resolution or other data but now it’s handled for sure.

The problem wasn’t related to my VPS setup but some feed addresses returning errors on the TT-RSS side. I once saw that some feeds were returning errors as their publisher made URL changes. Back then, I didn’t have the time to go through the problematic feeds but the idea struck me a few days prior that what if the problematic feeds are causing an update loop problem.

Once I corrected/removed the problematic feed URL, my feeds started updating as usual.

I assume there is a bug here and to be sure about it, I will try to replicate the issue in the coming days as I’m quite busy at the moment. Once done, if there is an I will issue an appropiate bug report.

If a feed isn’t working TT-RSS should move on to the next feed. If a broken feed is causing the updater to have a fatal error and abort that shouldn’t happen. If one or more feeds were causing a fatal error can you share them here so the cause can be identified and fixed? If you want to keep things private just send them as a private message to fox or me.

No, it’s fine; I can share them here. It’s just that I wanted to be able to reproduce the issue if possible to have something more concrete and I was thinking of creating a separate bug report for a better future clarity. But anyways.

It’s a bummer on my side that I didn’t note the problematic feeds before hadnling them. Although, I managed to track 2 out of 4 of them:

I’ve checked out the log file(Logs begin at Sat 2020-01-25 14:03:00 CET, end at Thu 2020-02-13 09:31:21 CET) and the only error that caught my eye is:

Jan 25 14:05:00 vps update_daemon2.php[24115]: PHP Fatal error: Uncaught PDOException: SQLSTATE[HY000]: General error: 2006 MySQL server has gone away in /var/www/rss.cansurmeli.com/public_html/classes/rssutils.php:136

But, regardless, here is the full log if that helps. Hope, the file isn’t expired by the time you grab it.

I see that the above problem is a MySQL error and that you don’t directly support MySQL.

And, by the way, the above data in the previous messages related to my setup still apply.

I tried to be as through as I can be but if there are other things I can provide, do tell me please.

i should’ve dropped mysql support altogether years ago

protip: the abortion of a database server you’re using is crashing. figure why its doing that instead of blaming tt-rss.