dday61

Ok. Got that worked out:

comments: 0

:start
A simple post to start getting Postie stabilised and ready for rollout to a beta test user. When it's stable I'll set it up on the other site and roll it out to more copy creators. Uncertain, at this stage, whether:

– they will be given a form which sends an email to me for validation, or
– they will submit their copy direct to Postie to be published with draft status.
That will be determined as I learn of the pitfalls of allowing direct submission.
:end

Not sure it was what was intended in the design of Postie, but it works. I can live with it so I’ve marked this issue resolved.

More from the host’s support:

We have found one error from your wordpress site and its should be fixed by disabled zlib.output_compression on php.ini file in your account. and pasted the error message as the following:-

———————————-
Notice: ob_end_flush(): failed to send buffer of zlib output compression (0) in /home/ihqxtua/public_html/wp-includes/functions.php on line 3743
==================================

I now get the expected response from the Configuration Test. I have, therefore maked the issue as resolved.

Response from host’s support team:

Please note that these are the default email ports in the server we never blocked 110, 993, 995 in our shared servers. So you may please contact with the plugins support forum and provide the following error from your account.

——————————-
[07-Aug-2018 08:57:00 UTC] PHP Fatal error: Maximum execution time of 300 seconds exceeded in /home/xxxxxxx/public_html/wp-content/plugins/postie/postie.class.php on line 1653
[07-Aug-2018 08:57:00 UTC] PHP Warning: Error while sending QUERY packet. PID=515 in /home/xxxxxxx/public_html/wp-includes/wp-db.php on line 1924
================================

I specifically asked about anything that might need to be unblocked when I asked if they could shed any light on the problem. I’d be surprised if that didn’t prompt any thoughts.

On command line/terminal access I think I may have had a bit of a brain wobble and thought I needed terminal access at the actual server. Doh! After putting my brain back in gear I get:

local server/pop gives:

derek@derek-kubuntu:~$ telnet mail.XXXXXXXXXXX.com 110
Trying a.b.c.d…
Connected to XXXXXXXXXXX.com.
Escape character is ‘^]’.
+OK Dovecot ready.
quit
+OK Logging out
Connection closed by foreign host.

Local server/imap gives:

derek@derek-kubuntu:~$ telnet mail.XXXXXXXXXXX.com 143
Trying a.b.c.d…
Connected to XXXXXXXXXXX.com.
Escape character is ‘^]’.
* OK [CAPABILITY IMAP4rev1 LITERAL+ SASL-IR LOGIN-REFERRALS ID ENABLE IDLE NAMESPACE STARTTLS AUTH=PLAIN AUTH=LOGIN] Dovecot ready.
C logout
* BYE Logging out
C OK Logout completed.
Connection closed by foreign host.

For gmail I get pages full of connection and certificate information that all look ok to me.

The obvious difference (with the local server) is that it replies “Dovecot” instead of POP or IMAP.

By ‘local’ I mean the same host that the wordpress site is on – as far as I’m aware it might not even be on the same server. TBH I don’t really mind whether I get it working from the ‘local’ host or from a Gmail account. Having fiddled, for most of yesterday and today, with the Jetpack offering (‘Post by Email’) I just want to stay away from it. This is a site that has been hacked once already (in a former Joomla guise) and still attracts the ‘wrong’ sort of visitor. I’d like to rename it but can’t because of what it is. I prefer to keep xmlrpc.php access locked down.

I’ll pass your link on to the host’s support team and see if it sparks any thoughts.

Thanks for the response.

I have received the following from the host’s support team:

I have found the following error from the error logs ‘ Maximum execution time of 120 seconds exceeded’, so I increased the Maximum execution time to 300 for both your domain and subdomain. You can also edit this through the cPanel >> Software >> MultiPHP INI Editor. And please check it out and let us know if you will get any issues.

———————————
[03-Aug-2018 15:59:16 UTC] PHP Fatal error: Maximum execution time of 120 seconds exceeded in /home/ihqxtua/public_html/rebuild/wp-content/plugins/postie/postie.class.php on line 1653
=================================

This seems to relate to:

    function log_onscreen($data) {
        if (php_sapi_name() == "cli") {
            print( "$data\n");
        } else {
            //flush the buffers
            while (ob_get_level() > 0) {
                ob_end_flush();
            }
            print( "<pre>" . htmlspecialchars($data) . "</pre>\n");
        }
    }

Line 1653 is ob_end_flush();

I don’t know whether that helps narrow the problem down or not.

Changing the time-out made no real difference (except that the Configuration Test ran longer before it returned only the two header lines) – “Postie Settings (v1.9.24)” and “Postie Configuration Test”

Set up a subdomain with only a fresh install of WordPress and the Postie plugin.

Still no joy connecting to either a local (on the same server) email address OR a gmail address.

Any suggestions?

I tried setting up a connection on the second site (as mentioned above) and have no trouble establishing a connection to an email address on that server. I get a screen of php and WordPress config details followed by the ‘Clock’ details, email connection details and the correct # of waiting messages.

The problem therefore seems to be an error on rachaelrowe.com which is unhandled by Postie’s Configuration Test – making it difficult to debug and correct.

Might this be possible using a symlink?

Why would there be a ‘standard’ time? They’re not the sort of people who will stick to any standards or conventions. They know that if they sustain their attack – and with improvement in technology – they know they will get through eventually. Be sure to take regular backups – the frequency of which will depend on the purpose and update frequency of your site.

I had a sustained BF attack, on one of the sites I maintain, that lasted for over a month. It was a site that got successfully hacked when it was on Joomla about 5 years ago. After it was hacked it was resurrected as a WordPress site. The attacks continued so I assume the URL is on the dark web somewhere.

I considered changing the URL but as it’s the name of an international travel author that would have been difficult.

Malicious activity, until recently, peaked and troughed. It was only when I started banning suspicious IPs for month, and those who try to use a non-existent user-id permanently, that the malicious activity started to show a significant decline.

I use WordFence on my live site without the deactivation problem. I do not use it on my test system.

Again – might, or might not mean anything as they are both very complex plugins. Just another anecdote that might, or might not, lead somewhere worthwhile.

@sunkissmarc: I don’t really see it as a workaround as I know the purpose of the file I’m uploading before I upload it and, therefore, know which folder to upload it to. Why use two processes when one will do?

The structure I have is that some folders are for images that will be permanently retained whereas others are for ephemeral images used in Calendar events. If I need to use an image in both I upload a duplicate to the relevant folder. Once the events for (say) 2018/04 are over their records will be reused and will no longer link to the images in the events/2018/04 folder – which can be cleared down and removed without risk of the front end suffering missing image problems. MLF helps to stop me accidentally deleting permanent media files amongst the ephemeral ones.

You might not, of course, have that luxury. Many systems/users have different requirements.

You’re right that any plugin should be able to integrate with the core functionality gracefully. If MLF is causing problems the cause needs to be corrected if possible or, if not, properly documented.

@sunkissmarc: Once I installed MLF I stopped using the core process to upload files – and used the MLF file uploader. I also used ftp, followed by MLF’s SYNC functionality for the few bulk uploads I had to do. That might be why I’m not seeing the same problems as you.

I wouldn’t know the calendar code to that depth. That’s something Time.ly themselves will have more knowledge of. If collected anecdotal experience can point to a possible cause I’m happy to share it.

I found that a local system I set up at home (on my laptop) for testing always deactivates the ai1ec posterboard extension when the server is powered down overnight. To reactivate it I have to log out of Time.ly, then back in again.

I don’t recall any problems at all with the posterboard extension/plugin on the live server. This tends to have excellent uptime.

Might be related. Might not. How’s the reliability of the server you’re having trouble with? Downtime? Generally good performance? Or frequent slow patches?

I recently started using the free version of Media Library Folders. I, too, had problems while implementing it and reorganising my media folders to improve the way the site works.

I don’t view the problems I had as ‘bugs’ in Media Library Folders – instead, I view them as a legacy left over from the early development of my site. I challenge any plug-in writer to be able to identify and write software that successfully corrects all the problems caused by other plugins.

During my site’s development I tried several media plugins, and they each left their own trail of damage and detritus. It then became the live system. Thanks to Media Library Folders the trail of damage and detritus is now gone, I have saved (off the top of my head) about 10 – 20% of the space my uploads folder occupied at the start of the exercise, and the media is now far more logically pigeonholed so I know what must remain, and what is ephemeral/transitory.

While doing this data cleansing exercise I took the trouble to look under the WordPress hood – as a result of which I have a better understanding of how WordPress works and, more to the point, where WordPress falls short in its handling of media. I now have enough confidence in my ability to write simple database/media validation scripts to warn me when media ‘goes missing’ and, therefore, will enable me to give my site’s visitors a better user experience. As with all good validation scripts they should be seperate from the primary data handlers (WordPress and Media Library Folders) and will, therefore, highlight possible problems in both – as well as problems introduced by my own working practices.

In my experience – Media Library Folders does not ‘mess up the wp media system’. Rather, it brings to light problems that existed before Media Library Folders was added to the system and allows the webmaster to correct them.