SuiteCRM 8.3 very slow - one page load ca. 81 seconds

Hello,

Iā€™ve installed several SuiteCRM instances since years on VPS and shared hosting and it was always acceptably quick.
On one new shared hosting environment Iā€™m running a small and fast website (Google Pagespeed 98 on mobile) and SuiteCRM 8.3 (new installation with 2 leads manually created).

  • One page load takes around 83 seconds
  • PHP 8.0
  • MySQL 8.0
  • opcache.enable true
  • max_memory_limit 512M
  • no 404 / 403, everything loads nicely - just very slow
  • suitecrm.log is almost empty
Wed Jun 28 06:56:03 2023 [2477233][1][FATAL] Mysqli_query failed.
Wed Jun 28 06:56:03 2023 [2477233][1][FATAL]  Query Failed: SELECT favorites.*
                  FROM favorites
                  JOIN config ON ( config.id = favorites.parent_id )
                  WHERE favorites.assigned_user_id = '1'
                    AND favorites.parent_type = 'Administration'
                    AND favorites.deleted = 0
                    AND config.deleted = 0
                  ORDER BY favorites.date_entered DESC LIMIT 0,10: MySQL error 1054: Unknown column 'config.deleted' in 'where clause'

Could that be related to a very slow system?

  • Iā€™ve moved the DB to my VPS (with another SuiteCRM running swiftly and connected to that one - but the speed is the same. VPS with 10GB RAM, very little traffic, MariaDB - and still the same issue.

On the shared host, there is a

  • Apache 2.4
  • NGINX reverse proxy
  • Caching enabled / disabled for testing purposes (all without changes / improvements)

The shared hosting company was quite helpful on the phone, but couldnā€™t see an issue as well.

In the browser, it seems like, most things load fast BUT the graphql endpoint:

Any idea, what that could be?

Iā€™m struggling with the same thing, with SuiteCRM 8.1. I have done the following so far:
(a) Checked and changed permissions of folders as required.
(b) removed errors that were showing the log
(c) Repair and rebuild
Nothing so far.

The server environment is perfect, I have other apps working on it fine. Here are some detials:

|Apache Version|2.4.57|
|MySQL Version|5.7.42|
|Architecture|x86_64|
|Operating System|linux|
|Shared IP Address|103.187.238.34|
|Path to Sendmail|/usr/sbin/sendmail|
|Path to Perl|/usr/bin/perl|
|Perl Version|5.16.3|
|Kernel Version|3.10.0-962.3.2.lve1.5.39.el7.x86_64|

The CPU, Memory etc are almost idle.

Enabling slow log does not show anything in the log. The loading animation shows for almost 10 second after each click - which is really bugging.

That MySQL FATAL is solved with this PR

It was merged 3 weeks ago and should be included in the next version.

1 Like

Thanks.
Do you think that could be the issue of the slow page load? Seems to be far fetched?

I have no idea. Try it. At least itā€™s always a good idea to remove FATAL errors, they stop script execution, so you never know what else is broken until you remove themā€¦

I have fixed all FATAL errors, there are none now. Yet, SuiteCRM is very slow. When I enable DEBUG level logs, I see this info:

Query Execution Time:9.9897384643555E-5
Query Execution Time:6.0081481933594E-5
Query Execution Time:7.7009201049805E-5

a basic query like:

Query:SELECT id FROM outbound_email WHERE type = ā€˜systemā€™ AND deleted = 0
Query Execution Time:6.9141387939453E-5

is this normal? almost all queries are above 6 seconds.

@anils please start your own thread, itā€™s difficult to help two people at once.

Run the first query in this post and copy your results into the new thread you create:

You probably have overgrown tables

1 Like

What happens at the /api/graphql endpoint?
What could it be, when it takes 20 and more seconds to load?
All CSS / JS load instantly, only the graphql takes very long?

@pgr Iā€™ve applied the patch - FATAL error is gone, performance is still bad.

Donā€™t know, some web server configuration thingā€¦ maybe related to the mod_rewrite.

I know that turning on opcache makes a big difference in terms of performance for v8, but it should never be taking 20 seconds in the first place, on a system thatā€™s not loaded with activity from multiple users, so thereā€™s probably something else going on here.

The hosting company is working on it.
Opcache was switched on already.

In the meantime Iā€™m using a VPS.