PS: I was able to grab these from the SP Admin Error Log… sorry for the post length and poor formatting, but I thought something here may help.

Date    #    Type    Description 31 October 2013
4:52 pm    2    php    file: s/simple-press/forum/content/sp-search-view-functions.php
line: 61
function: sp_SearchHeaderName
Notice | Undefined variable: label

31 October 2013 3:12 pm    1    php    file: s/simple-press/sp-api/sp-api-timezone.php
line: 53
function: sp_apply_timezone
Notice | Undefined index: timezone

31 October 2013 2:36 pm    9    php    file: ripawds/public_html/wp-content/sp-resources/forum-plugins/admin-bar/library/sp-admin-bar-components.php
line: 564
function: sp_AdminBarDashboardPosts
Notice | Undefined property: stdClass::$topic_id

31 October 2013 2:36 pm    9    php    file: ripawds/public_html/wp-content/sp-resources/forum-plugins/admin-bar/library/sp-admin-bar-components.php
line: 562
function: sp_AdminBarDashboardPosts
Notice | Undefined property: stdClass::$post_count

31 October 2013 2:36 pm    10    php    file: ripawds/public_html/wp-content/sp-resources/forum-plugins/admin-bar/library/sp-admin-bar-components.php
line: 559
function: sp_AdminBarDashboardPosts
Notice | Undefined property: stdClass::$post_count

31 October 2013 5:17 pm    25    php    file: s/simple-press/forum/content/sp-common-view-functions.php
line: 1549
function: sp_BreadCrumbs
Notice | Undefined index: forumname

31 October 2013 5:18 pm    127    php    file: s/simple-press/sp-startup/forum/sp-forum-support-functions.php
line: 635
function: sp_setup_title
Notice | Undefined index: forumname

1 November 2013 2:00 pm    4    php    file: s/simple-press/forum/content/sp-common-view-functions.php
line: 1555
function: sp_BreadCrumbs
Notice | Undefined index: topicname

31 October 2013 4:18 pm    9    php    file: s/simple-press/sp-api/sp-api-filters.php
line: 1414
function: sp_format_display_image
Notice | Undefined offset: 1

1 November 2013 2:00 pm    45    php    file: s/simple-press/sp-startup/forum/sp-forum-support-functions.php
line: 639
function: sp_setup_title
Notice | Undefined index: topicname

31 October 2013 4:16 pm    13    php    file: s/simple-press/forum/content/classes/sp-search-view-class.php
line: 255
function: sp_build_search_url
Notice | Undefined index: forum

31 October 2013 3:25 pm    8    php    file: s/simple-press/sp-startup/forum/sp-forum-support-functions.php
line: 85
function: sp_populate_query_vars
Notice | Undefined index: forum

31 October 2013 4:55 pm    22    php    file: s/simple-press/forum/content/classes/sp-search-view-class.php
line: 81
function: sp_searchview_control
Notice | Undefined index: forum_id

31 October 2013 11:54 am    4    php    file: s/simple-press/admin/panel-options/support/spa-options-save.php
line: 37
function: spa_save_global_data
Notice | Undefined index: blockredirect

31 October 2013 11:54 am    6    php    file: s/simple-press/admin/panel-options/support/spa-options-prepare.php
line: 30
function: spa_get_global_data
Notice | Undefined index: blockroles

1 November 2013 2:00 pm    127    php    file: ripawds/public_html/wp-content/sp-resources/forum-plugins/private-messaging/library/sp-pm-components.php
line: 218
function: sp_pm_do_remove_pms
Notice | Undefined index: sfpmremove

you can ignore those notices – they are harmless and we fix them as we come across them… starting with 5.3.3 (or 5.3.4) you will see more of them since we found we had actually left some test code in place that was omitting the undefined index notices…  so they were re-enabled which just means sifting through them….  many have already been fixed in 5.3.5…

I can think of no reason why your sfposts table would not have any indexes…  every site when installed creates the tables and the needed indexes… so unless your mysql permissions do not allow adding indexes, unknown how it could happen..  just looked at my couple of multisite test sites and they all have proper indexes on the tables…

yes, you can add them via phpmyadmin…  size should be a matter… I have added them on a site where the table size was over 2 GB…

here is a screen shot of what the standard indexes look like on the sfposts table…  I have not compared those to what the install script does, but should be quite close…

Mr Papa said
here is a screen shot of what the standard indexes look like on the sfposts table… 

Interesting. We do have indexes on our wp_1_sfposts tables as shown in the first screenshot below, very similar to the defaults you provide. I do, however, have an additional table and get the following notice:

The indexes user_id_idx and user_idx seem to be equal and one of them could possibly be removed.

Before I do more harm than good, my questions at this point include:

• Could the “equal” user_id_idx  index be causing problems and could I safely drop it? I never added that so i have no idea how it was created.
• Is there a way to rebuild or repair these indexes? I see no commands in PHP MyAdmin other than Edit and Drop.
• Admittedly, I don’t understand what “Cardinality” means in this view but it appears the other numbers refer to forums, topics, comments, etc. So should post_content really be 1 when there are 110,317 posts?
• Anything else look odd in the attached Indexes and Table Info screens? Is 53+ MB normal for the index on a table this size?

Any other suggestions as to what might be causing MySQL resources to climb immediately upon SimplePress activation? Thanks again for any input on this!

I am going to have to pass on why you have two indices on user_id but yes – I would remove the one named user_id_idx. That will be fine.

Do you regularly optimise your tables? This action will normally correct any indexing issues as well as ensure the table data structures are correct etc. Normally, at the bottom of the table listing in phpmyadmin there is a drop down list of options and from there you can ‘check’, ‘repair’ and ‘optimise’ tables. I would suggest you perform a ‘check’ operation n all of them to see if anything gets reported as a problem. And then repair and optimise as necessary. This should also rebuild the post_content full text index as well – which should not be empty.

Yellow Swordfish said
…I would remove the one named user_id_idx.

Thanks, done.

Do you regularly optimise your tables?

Yes, about every week or so I do regular checks and optmize db tables.

I just ran a check on all sf tables with OK results all around. After deleting the duplicate index, I then ran a repair and optimize on each sf table. Again, Status returned was OK for all tables.

So… I activated SimplePress and the server load quickly climbed to 5.5±, but stabilized long enough for me to unlock the forums and run the Rebuild Indexes command on all forums from the SP Admin. Then, within a couple minutes the load climbed to 20± before I deactivated the plugin. Load then immediately dropped to <0.5.

During this short time, we did receive about 10 guest spam posts. To rule out some sort of forum based spam attack, I waited until loads settled down and reactivated SimplePress. I immediately locked the forums, but continued to watch loads climb to 15+ within seconds. Loads again immediately dropped upon deactivation of the plugin.

FYI: I have ruled out any plugin conflicts by deactivating all other but SimplePress, and same problem persists.

This should also rebuild the post_content full text index as well – which should not be empty.

I realize this should not be empty but I wonder if a Cardinality value of 1 is appropriate for forums our size or if this may be an indication of something wrong.

What steps can I take from here to get our forums back online and stable?

Is there a recommended collation for _sf tables? i may be reaching for straws here, but I’m at my wit’s end and Tripawds members are getting restless.

Jim – could you check your php error log? Not the SP one – the actual, full php log?

Yellow Swordfish said
…could you check your php error log?

I’ve been monitoring the server error log via cPanel in hopes that it would shed some light, but nothing unusual shows up other than a lot of denied access requests because of our firewall settings.

There are no SQL related errors in the log, and the only mentions of “simple-press” include…

[Sun Nov 03 12:27:12 2013] [error] [client 67.142.130.48] File does not exist: /public_html/wp-admin/images/ed-bg-vs.gif, referer: …/wp-admin/admin.php?page=simple-press/admin/panel-options/spa-options.php

This one is interesting since it doesn’t appear that image is in the core wordpress download.

[Sun Nov 03 12:27:09 2013] [error] [client 67.142.130.48] File does not exist: /public_html/wp-content/plugins/jetpack-lite/_inc/images/menuicon-sprite.png , referer: …/wp-admin/admin.php?page=simple-press/admin/panel-options/spa-options.php

And this one is interesting since that image file does exist, though it looks like there is a blank space after the filename.

<still perplexed>

OK, this may reveal a bit more…

I noticed the physical error_log file in our root directory had grown so extraordinarily large that I couldn’t view it. So I regenerated the log file, reactivated SimplePress and kept it on for a few minutes this time before I started noticing the loads increase again.

I checked that log file and it’s filled with errors. Could you or Steve please review this full error_log file? here are a few interesting excerpts….

[03-Nov-2013 20:10:11 UTC] PHP Notice:  Undefined index: forumname

[03-Nov-2013 20:15:53 UTC] WordPress database error MySQL server has gone away for query UPDATE `wp_1_sfoptions` SET `option_value` = ‘a:4:{s:16:”css-sp-theme-css”;a:2:{s:8:”modified”;i:1378753824;s:4:”type”;s:3:”all”;}s:3:”ver”;a:1:{s:3:”all”;i:61;}s:16:”css-sp-admin-bar”;a:2:{s:8:”modified”;i:1378299212;s:4:”type”;s:3:”all”;}s:17:”css-sp-share-this”;a:2:{s:8:”modified”;i:1356538920;s:4:”type”;s:3:”all”;}}’ WHERE `option_name` = ‘sp_css_concat’ made by require(‘wp-blog-header.php’), require_once(‘wp-includes/template-loader.php’), include(‘/themes/wpmu-nelo-child/page-full-width-forums.php’), get_header, locate_template, load_template, require_once(‘/themes/wpmu-nelo-child/problem-with-post-edit-buttoneader.php’), wp_head, do_action(‘wp_head’), call_user_func_array, wp_print_styles, do_action(‘wp_print_styles’), call_user_func_array, sp_load_plugin_styles, sp_combine_plugin_css_files, sp_update_option

[03-Nov-2013 20:16:22 UTC] WordPress database error MySQL server has gone away for query SELECT  * FROM wp_1_sftopics WHERE topic_slug=’chemo-for-osteosarcoma-yes-or-no’ made by require(‘wp-blog-header.php’), require_once(‘wp-includes/template-loader.php’), include(‘/themes/wpmu-nelo-child/page-full-width-forums.php’), get_header, locate_template, load_template, require_once(‘/themes/wpmu-nelo-child/problem-with-post-edit-buttoneader.php’), wp_head, do_action(‘wp_head’), call_user_func_array, sp_forum_header, sp_setup_meta_tags, sp_get_metadescription, spdb_table, spdb_select

[03-Nov-2013 20:16:22 UTC] WordPress database error MySQL server has gone away for query SHOW TABLES LIKE ‘wp_1_sferrorlog’ made by require(‘wp-blog-header.php’), require_once(‘wp-includes/template-loader.php’), include(‘/themes/wpmu-nelo-child/page-full-width-forums.php’), get_header, locate_template, load_template, require_once(‘/themes/wpmu-nelo-child/problem-with-post-edit-buttoneader.php’), wp_head, do_action(‘wp_head’), call_user_func_array, sp_forum_header, sp_setup_meta_tags, sp_get_metadescription, spdb_table, spdb_select, sp_construct_database_error

“MySQL server has gone away”  

Not sure what this means as our sites are still operating and the database is on the same server.

Hopefully something here or in all these errors may help explain where I can go from here.

THANK YOU!

The ‘server gone away’ message can be a hard one to track down. And the problem with it is that once it has gone away it can spawn a lot of error entries..

It is probably best to just ask you to do a search on the term as there is a lot of documentation on this available but basically it means your MySQL Server is timing out and stopping. The difficult question is… why. And the secondary question is why has it started now. You are probably going to need to involve your host support in this one. But some questions…. Has your MySQL Server been updated recently? Reading suggests that some versions can be more prone to this than others. Has the timeout value been altered for any reason? Have any MySQL settings been changed in the server config of late? Actually even the Apache settings may have an effect.

Yellow Swordfish said
It is probably best to just ask you to do a search on the term…
But some questions….

Thanks Andy, I have been frantically researching this, and been in close contact with our server manager. We have been running the latest version of MySQL that is compatible with cPanel, and have been for some time.

The only recent changes on our site and the server were my updates last week of both SimplePress and WordPress, on the same day that we started noticing the crippling MySQL issues.

With SimplePress deactivated, all is fine. I’m hoping someone here might be able to provide some direction for further troubleshooting, or recommendations for appropriate MySQL / Apache / server config settings.

Steve? Anyone?

After further discussion with our server manager, I’ve confirmed that nothing has changed on the server side.

He says…

I looked at the mysql log file and didn’t see anything that stands out in recent records. My understanding is we have done all this tweaking of the conf files and, up until you did the plugin upgrade everything was running fine. And disabling the plugin made everything run fine again. So, we could futz around forever trying to find some setting that the new version of the plugin in relationship to your database might require, but given there are so many variables, I am hoping someone who has tread these waters before might give us some clues.

So with that in mind, I asked for our current my.conf file. Perhaps someone could review these settings and/or recommend any suggested changes…

Steve made some changes to ours about a year back when we saw some of the ‘gone away’ messages on our server so he may recognise something perhaps. he will be back tomorrow.

I did check the last couple of updates and I see no real changes to the general mysql queries. Minor updates is all. Perhaps Steve will recognise something or have an idea.

Thanks again, I’m just going crazy trying to get somewhere on this…

As I understand things you usually in the past have updated a test rig first. Did that biot hapen this time? And if it did does that show the same issues?
Perhaps you could also conform the version you were using prior to the latest update. I can check code changes again although truth is we have no other reports at the moment of this behaviour.