very urgent issue with recurrence instances

[Steve]

I’ve found through experience that if you add to threads here to ask why there’s been no response, your thread simply goes to the bottom of the queue again and you have to wait longer. So I can’t think of anything else to do apart from start a new thread.

My issue is here:

Sudden increase in number of events

I’ve been trying to resolve this all week. The client is very anxious because the crucial part of their site, the calendar, is now crashing regularly at a peak use time.

I’ve made some progress in diagnosing this with your engineer, but response times have been appallingly slow. It’s just not acceptable for a paid product which is crashing the site.

Let me summarise my current position:

1. There are two events which have about 37,000 and 14,000 recurrence instances in the posts table (with corresponding meta). The next highest is about 120, so clearly these posts are the problem.
2. I’ve examined the recurrence rules (screenshots attached). The engineer on the other thread suggested it might be badly-configured rules, but I can’t see anything wrong there.
3. I’ve examined the meta data for the recurrences for these events (specifically the event start date). None of the start dates is outside the range specified, which isn’t very far in the future. I have the system set to create for 12 months in the future, but all the recurrence instances respect the ‘until’ date specified, which is about 1 month in the future.
4. It seems clear this is a bug in the plugin, almost certainly related to your recent update which had problems with recurrences.
5. I’ve tried the plugin provided to fix this, which runs a WP cron job. It didn’t seem to be doing anything so I was pointed to another plugin which has a ‘rebuild recurrence’ link for each event. I’ve tried this but I just get a white screen each time I click this for the 2 problem events in question. When I reload the page and check the database, nothing has happened.
6. I’m on the verge of simply deleting all recurrences via SQL, or maybe just deleting the events. However, I’m not sure if such a blunt approach might cause other issues, and I’m keen to not make a problem worse.
7. I’m very disappointed with the poor support response on this, and would appreciate it if you could address this properly.

Thanks.

[Brook]

Howdy Steve,

I would love to help you with this. I have read through your old post and reached out to Cliff to let him know there is a new response here.

I’ve tried the plugin provided to fix this, which runs a WP cron job. It didn’t seem to be doing anything so I was pointed to another plugin which has a ‘rebuild recurrence’ link for each event. I’ve tried this but I just get a white screen each time I click this for the 2 problem events in question. When I reload the page and check the database, nothing has happened.

Based on your description in the other topic it sounds like both of those plugins were actually working, perhaps working too hard. The white screen is an indication that they are cleaning things in the background, but the speed at which they are cleaning them is too fast for your server’s configuration.

You mentioned with the original plugin, which you deleted, that it would intermittently show 500s. Again this is a strong indication that the plugin was working, but again probably too fast. Ironically even at it’s current speed I would not be surprised if it took a couple of days for it to finish cleaning everything in your database.

I’m on the verge of simply deleting all recurrences via SQL, or maybe just deleting the events. However, I’m not sure if such a blunt approach might cause other issues, and I’m keen to not make a problem worse.

If you ask me this is a great idea! There is actually pretty simple SQL query to delete all recurrences from an event. It sounds like if you run it twice, once for each event series, it will cleanup nearly 30000+ entries. That’s not shabby at all. Best of all since you will be running it as a SQL query rather than through WordPress, if it errors out it will give you an error and let you know why.

Here is the query for rapidly delete all recurrences of a parent event:

https://gist.github.com/elimn/e5de5d2037678de0b926c307022571c5

If you walk through the steps outlined in the comments there, backing up your database, getting the parent ID, and executing the query, does that cut down significantly on the # of events?

I’ve found through experience that if you add to threads here to ask why there’s been no response, your thread simply goes to the bottom of the queue again and you have to wait longer. So I can’t think of anything else to do apart from start a new thread.

Yeah that is very true. Like most WordPress support forums our community is modeled after the WordPress.org forum. Thus our forum software, much like that on .org, will assume that a bumped topic is someone responding that their issue has been resolved. If you need to update an issue the Edit button should be used. This is great for consistency and preventing the need to reread the rules of each forum, but sometimes it’s not what a given person wants.

I’m very disappointed with the poor support response on this, and would appreciate it if you could address this properly.

I totally get where you are coming from. We are almost all freelancers on the side here at Modern Tribe, and have had sites go down our selves. I’m sorry this is being such a pain. The main reason I wanted to address this is to set the expectation at the outset that we can frequently only respond once every 24 hours. Cliff has on average been responding in a couple of hours, unless the post happened at night in which case he comes back early the next the morning. That is going way above and beyond the support agreement you signed, and honestly it is going to nigh impossible for us to top or even maintain his current response times. We do our best to give clients like you who are having serious issues extra fast response time, but Cliff has already been doing exceptional in this area already. He is exceeding our 24 hour typical time often by 10x. I am truly sorry that this is not fast enough for you and I definitely understand why it’s not, but it is simply not possible for us to do better. If you keep opening up duplicate topics it is very possible response times will actually decrease as we will have to coordinate behind the scenes between each person who is helping you, in order to determine where each of us is heading with our line of thinking. Again I am sorry this is not what you want, but it is the reality.

Please let me know how the above goes. Does that SQL query help delete things far more rapidly?

• Brook

[Steve]

Thanks for getting back. Appreciate the extra effort, I guess I’m falling foul of time zones? What time zone are you in? Also, do you provide a premium service at all? I’ve had a few problems with this plugin that have necessitated back-and-forth, and it’s just not realistic for my client to have this level of backup when it’s a plugin-specific issue affecting the site.

Anyway:

1. The excess posts mean that I can’t actually back up the database via phpMyAdmin. I’m going to have to wait until the server admins are there in the morning so they can do a backup. The SQL looks good but obviously I can’t run any SQL straight onto a live DB without backing up first!
2. I tried the SELECT with the event that has 14,071 recurrences, and there were 408,037 results. Is that because of the meta and taxonomy joins (which are presumably to clean those bits up with the posts)? Assume so, just thought I’d check.

[Brook]

Thanks for getting back. Appreciate the extra effort

You are very welcome. I’ll pass that on to Cliff as well.

Early on in my career I had a server with ~30 sites get hacked, and I would not wish that type of stress upon anyone. Whatever I can do to help I am here for.

I guess I’m falling foul of time zones? What time zone are you in?

Personally I am in GMT -7. But we are not all in the same spot.

Also, do you provide a premium service at all?

We do not. We have repeatedly investigated this notion and each time found it to be not quite viable. If you have any feedback you wish me to pass on here, or if you want to talk to the gentleman who heads up support, just let me know.

I tried the SELECT with the event that has 14,071 recurrences, and there were 408,037 results. Is that because of the meta and taxonomy joins (which are presumably to clean those bits up with the posts)? Assume so, just thought I’d check.

Exactly. That query is essentially a combination of the post meta table, the post table, and the taxonomy one. It is a simple way of selecting/deleting all standard WordPress data for a given set of posts.

It is much much faster at deleting posts than the WP API. However it is typically a good idea to use the WP API, just in case any third party plugin has attached nonstandard data or routines to the WP delete functions. Our cleanup plugins uses the API, and thus runs a lot slower. But running this query is usually completely safe, and so when you’re in a pinch like this it is a worthy option. If something does not work right after the query, well that’s what your backup is for. I have run this SQL query myself many times and yet to experience odd behavior, hence why I am recommending this as the preferable option for you.

Please let me know if you have anymore questions in the mean time. I will be logging off soon but I will do my best to check in once more before tomorrow.

Cheers!

– Brook

[Brook]

Howdy Steve,

I just wanted to check back in. I’ve been hoping that no news is good news. But if you do need any more help please let me know.

– Brook

[Steve]

Thanks, didn’t realise I hadn’t closed this thread! The SQL worked a treat – all’s well that ends well.

[Brook]

That’s no problem! I am happy to hear the SQL worked. Thanks for getting back.

Let us know if you ever need anything else. Hopefully no more bugs though aye?

– Brook

[Author]

Posts