Support community for TTG plugins and products.
NOTICE
The Turning Gate's Community has moved to a new home, at https://discourse.theturninggate.net.
This forum is now closed, and exists here as a read-only archive.
You are not logged in.
Hi all,
update seems to have worked well so far.
Compliments for that and thank You for the comtinuing support and improvement.
After first viewing the updated site, I noticed that an external ressource is being blocked by the noScript plugin and Privacy Badger.
Fontawesome may not be the most evil party on the web, but I have always made a point of avoiding inclusion of content from external sources. (With re Captcha2 on the contact form being the only exception yet.) Proposing to my visitors to allow execution of scripts from external sites to view my site properly is something I don't feel comfortable with.
I have always liked the way, CE and BL handle the sharing with social media for the reason that no Like Buttons and such is being fetched externally.
Is there a chance to include the fontawesome stuff from a local path on my site like it used to be in BL2?
It's not a priority theme for me, but might make me feel better somewhen in the future.
Health and happiness to all in this community
Michael
Hi Oliver,
I guess that is content You entered in the fields "fb:admins" and "fb:app_id" in the advanced settings.
Have You tried entering the Umlauts as HTML entities? (ä Ä etc)
I like the pages a lot btw. They make Axel Kreisel appear to me as an artist (and a craftsman obviously) whom I would like to know more about. The socratic bodies are fascinating me.
Cheers Michael
Hi Oliver,
if I get You right, the problem concerns only the display of umlauts in search results.
Can You provide an example search stirng?
Did You set the language code at settings->languages to "German - de"?
Did You provide a Google Site Verification ID in BL settings?
I just switched from 7.2 to 7.3 because the recent WP update recommended to abandon 7.2.
7.4 is also available at my hosting service.
Hi all,
When I view a single blogpost page I get this between the post title and the post main copy:
Warning: DOMDocument::loadHTML(): Tag figure invalid in Entity, line: 6 in /www/htdocs/w0189dbd/kannznichkaufen.de/blog/wp-content/themes/pangolin-template/inc/theme_posts.php on line 179
Warning: DOMDocument::loadHTML(): Tag figcaption invalid in Entity, line: 6 in /www/htdocs/w0189dbd/kannznichkaufen.de/blog/wp-content/themes/pangolin-template/inc/theme_posts.php on line 179
Does this have to do with the new lazy loading feature, or is there something else I can adress?
A sample to check live is here:
https://kannznichkaufen.de/blog/tsg-spr … hausen-22/
Thanks a lot in advance.
Michael
I have a similar situation.
My Gallery with a download enabled template is https://kannznichkaufen.de/gal-TSG/sais … /index.php
When I select the download icon in the single view of https://kannznichkaufen.de/gal-TSG/sais … #gallery-2 I get an error. The URL that appears in the adress field of my browser is
https://kannznichkaufen.de/gal-TSG/sais … MG2462.jpg
When I edit this manually by deleting "/download.php?file=photos-for-download" the remaining
https://kannznichkaufen.de/gal-TSG/sais … MG2462.jpg gets the corresponding larger rendition is found.
I tried it with php versions 7.1 and 7.2, cleared and emptied caches.
Hi Daniel,
I wish it had been so easy for me.
I had first tried to copy all the content. The new provider even has a tool for that.
But my .htaccess files did not work well with the new server.
I had made some changes when I switched from http to https on the old domain.
And trying to fix that I messed up completly
Also, the new site has a different domain name too and the old site has been quite a mess after moving from ce3 to ce4 and then to BL.
So I have already done a fresh install of BL2 and transferred all my templates via ex- and import.
So now it's "only" the albums left.
First of all I want to wish everybody a fine christmas!
I have moved to another provider.
I have the new site running with another domain name and did a fresh install.
What is the easiest way to transfer the galleries published on the old sit to the new one?
Is there a method that saves me from creating all albums and album sets under a new publishing service in LR?
I came up with the idea to modify the publishing service instead of creating a second one.
I put the paths and API key of the new server and domain in the settings of the old publisher instance.
LR prompts me, if I would like to republish everything and I said "No".
When I touch the first of my albums to edit them, I get a message from LR/pubisher that the album does not exist on the server and that I should edit and republish.
This works with albums.
With album sets it's a bit different:
When I edit, publish and view in a browser, everything seems fine first.
But when I want to edit the same album set again and republish it, LR publisher again says the the album could not be found for a given id and that I should better edit and try again.
The changes I made are saved and can be viewed in the browser, but the id for the album set assumed by LR/publisher is different from the id used for the same album in BL2's publisher section.
If someone has an idea or good arguments to talk me out of being lazy, You're welcome.
All the best to You again
Michael
Yes. Exporting to WP worked alright. It's just the saving to Backlight that results in this message. Also it seems as if despite the error message, the items I changed are being saved correctly.
After I imported the template settings that I exported from my old site I was warned, that this templated had been created by an older version of BL and that I should better open it and the save again. This was when I first saw the message.
As I started writing this, I tried to reproduce again. the message is gone.
I have WP-Module installed and want to create my first Theme Template.
When I click save, I get "Unexpected error: Undefined variable: backgroundURL in background.php on line 71"
I did a reinstall of all modules. Deleted the template (which was originally imported from ma old site) created a blank one, but get the same results.
Hi again,
I finally have an answer from Hosteurope. They have not set a separate space quota for the user_tmp. It is accounted on my regular quota. Since > 10 GB of my Webspace are free, space should not be the cause of the problem.
They put in place a test script that allows me to verifiy that my user tmp is indeed writeable and also that it is empty. No left over stuff cluttering around.
I had proposed them to point the environment variable TEMPDIR which SQLite uses to find a place for temporary stuff to a location within my webspace where I could monitor what is going on there.
They refused that and proposed I should change to a dedicated Server or vServer.
@Ben: I would be glad if You could point out what steps it takes to switch back to Backlight 1:
Restoring the original content of the backlight directory is trivial. I already did that.
Which of the Database Backups must be reinstalled and what else has to be done to rollback the steps performed by the installer?
I want ot rollback and regret having applied the installer on my live site without testing.
@Alfred: Hosteurope weren't service oriented in my perception. They also weren't impressed by the fact that I'm not their only customer experiencing the problem. I have not yet decided what is best: switching to a virtual server or just trying another hoster. I feel like trying another shared hosting offer by all-inkl with a test domain first. I will keep You updated.
Cheers for now
Michael
Hi again,
Ben's question has something to it. My research also pointed in the direction of diskspace.
The user_tmp path is outside of the root I can access via ftp.
It might well be that ther is a separate quota that isn't shown in the customer admin panel.
Our hoster has a script that is claimed to show the content of the TEMPDIR.
That script gives no result as if the user_tmp where empty.
This may be true, but it can also be that the script doesn't work due to changes in permission settings. No way to verify.
I tried contacting the hoster with the question, if they could check my user_tmp.
First phone call was frustrating. They pointed me to the error logs, where I have masses of missing file errors. But they are due to the fact that I moved the old content of the Backlight folder out of the way.
Alfred: As soon as they send me a ticket number, I will tell them, that You have the problem too.
Let's hope it can be solved without their assistance.
Hi Ben,
55% of my quota is unused.
I read something about a transaction. Can it have to do with journal settings?
When I open the log in backlight/data/admin/logs. I don't read about the I/O error. But in my browser I see both.
[14-Sep-2018 10:23:25] ERROR: SQLSTATE[HY000]: General error: 1 cannot rollback - no transaction is active|#0 /is/htdocs/wp1139230_3D4UV0VX43/www/fotos/backlight/modules/module-framework/helpers/GenericUpgradeHelper.php(111): PDO->rollBack()#1 /is/htdocs/wp1139230_3D4UV0VX43/www/fotos/backlight/modules/module-framework/helpers/GenericApplicationHelper.php(138): GenericUpgradeHelper->upgradeIfNeeded()#2 /is/htdocs/wp1139230_3D4UV0VX43/www/fotos/backlight/modules/module-framework/Framework.php(440): GenericApplicationHelper->checkUpgrade(Array)#3 /is/htdocs/wp1139230_3D4UV0VX43/www/fotos/backlight/modules/module-framework/Framework.php(151): Framework->initApplication()#4 /is/htdocs/wp1139230_3D4UV0VX43/www/fotos/backlight/modules/module-framework/Framework.php(41): Framework->init()#5 /is/htdocs/wp1139230_3D4UV0VX43/www/fotos/backlight/index.php(31): Framework->__construct('..', Array)#6 {main}|URL: /backlight/index.php
Hi,
sorry, doesn't help. I already was at 1.2.2.
To do as Ben proposed, I must log into backlight. Whenever I try this, I get the output as above.
My PHP version is 7.1.
The provider announced to switch off 5.6 soon.
The ftp-account I created for Ben is still active. credentials in my mail from June, 7 2018.
Michael
Hello all,
I just started the Backlight installer. It seemed successful until I clicked "Proceed to Backlight".
After entering my credentials the database update started but seemingly did not finish.
I got this output:
Updating TTG Backlight from version 1.2.2 to 2.0.0
Backing up databases...
Backed up master.sq3
Upgrading to version 1.2.3...
Update to 1.2.3 completed
Upgrading to version 1.9.1...
Updating settings
Update to 1.9.1 completed
Upgrading to version 1.9.2...
Migrating settings
Unable to upgrade to version 1.9.2
Error message: SQLSTATE[HY000]: General error: 10 disk I/O errorSQLSTATE[HY000]: General error: 1 cannot rollback - no transaction is active|#0 /is/htdocs/wp1139230_3D4UV0VX43/www/fotos/backlight/modules/module-framework/helpers/GenericUpgradeHelper.php(111): PDO->rollBack() #1 /is/htdocs/wp1139230_3D4UV0VX43/www/fotos/backlight/modules/module-framework/helpers/GenericApplicationHelper.php(138): GenericUpgradeHelper->upgradeIfNeeded() #2 /is/htdocs/wp1139230_3D4UV0VX43/www/fotos/backlight/modules/module-framework/Framework.php(440): GenericApplicationHelper->checkUpgrade(Array) #3 /is/htdocs/wp1139230_3D4UV0VX43/www/fotos/backlight/modules/module-framework/Framework.php(151): Framework->initApplication() #4 /is/htdocs/wp1139230_3D4UV0VX43/www/fotos/backlight/modules/module-framework/Framework.php(41): Framework->init() #5 /is/htdocs/wp1139230_3D4UV0VX43/www/fotos/backlight/index.php(31): Framework->__construct('..', Array) #6 {main}|URL: /backlight/index.php
That's beyond my scope, I'm afraid.
As Ben might remember, there had been an SQLite issue with my provider recently. The essence was that SQLite3 did not have access to the temporary directory it tried to write to. The provider helped in the end by reluctantly setting the environment variable TMPDIR to a path that is writeable. According to phpinfo() this setting is still active.
The long story with all the try and error is here:
http://community.theturninggate.net/vie … hp?id=8723
Thank You in advance for Your help.
Michael
Quote from https://www.sqlite.org/tempfiles.html
5. Temporary File Storage Locations The directory or folder in which temporary files are created is determined by the OS-specific VFS.
On unix-like systems, directories are searched in the following order:
The directory set by PRAGMA temp_store_directory or by the sqlite3_temp_directory global variable
The SQLITE_TMPDIR environment variable
The TMPDIR environment variable
/var/tmp
/usr/tmp
/tmp
The current working directory (".")
The first of the above that is found to exist and have the write and execute bits set is used. The final "." fallback is important for some applications that use SQLite inside of chroot jails that do not have the standard temporary file locations available.
That reminds me of an issue I had with my provider after they had been acquired by GoDaddy.
It turned out that the provider changed the setup up in a way that sqlite tried to write te,porary stuff to a path it did not have permission for. (/var/tmp/). My stuff is on shared hosting and customer scripts should not be allowed to write outside the customers home directory. SQLite3 has a sequence of possible locations for temporary stuff. Setting the environment variable $TMPDIR to an apropriate path might help:
putenv("TMPDIR=/some/path/within/userspace/");
In my case this had to be done by the provider's support desk.
Wish You success.
Full Thread is here: http://community.theturninggate.net/vie … hp?id=8723
You can of course acheive this by manually packing and uploading a zip package to the server and including the URL of that file in the description of the gallery. You can use markdown syntax to highlight the link.
Hi Ben,
deleting the additional galleries went well. I downloaded and inspected publisher's database. There was no further debris left from my numerous test uploads.
The earlier proposal of the providers support to add some "sleep" to sqlite calls might be a hint that they are harmonizing standards to facilitate synergies.
I have always been content with them in the past. But in the past it was easier to reach a competent person on the telephone hotline. I liked being able to listen to and talk with people who are about to help me with problems.
That was why I have paid above average price for a hosting service for more than 15 years. They seem to have given away that distinguishing feature they had in the past. I have high respect for standardized and well documentetd processes. Having experts spend parts of their time in smalltalk and kindness isn't the route a profitable business should take. But in my case that led to a problem has not been handled. That signal "we screen our valued experts, don't call us, wee call you" made me guess I should better take the time to prepare the written information well and be precise.
in the end insisting that running a service as common as sqlite according to spec should be regarded as an exception and an act of courtesy gives me a strange feeling.
I had recommendations for all-inkl in the past from a friend who hosts multiple high traffic sites based on wordpress and phpBB there. But he does it on dedicated servers instead of a shared hosting enviroment, which will be configured differently. It would half my cost from the start. I could use a let's encrypt certificate instead of the one I now pay for. This means further savings when the current one expires. In a year or two I will run out of space. With HE I would have to switch to another contract AND to another physical server which had been some work and configuration trouble the last time I did that. They offer test accounts at all-inkl. I will get one as soon as work gets less ;-)
I still have some time to decide.
A fine weekend again.
Sent from the garden
Michael
Hi again,
it seems to be sorted. Template cache could be cleared, upload succeeded.
Hosteurope made the same changes to my configs they did to Alfred's. They also did a good job in having me understand that this was a matter of courtesy. I learned that they had been acquired by GoDaddy in 2016.
Ben, do You by chance recall if there are Backlight users successfully hosting with all-inkl.com ? Just asking for a friend.
I have some questions left:
All the interrupted attempts to publish the last album have left some directories that I will have to delete.
After deleting them in Backlight, is there anything I can do to give the database some care?
I had the value for "Delay API requests" in the settings of my publisher service put to a hight value, which had not been necessary before. What are typical values that help with GoDaddy?
Thank You for help, advice, patience and all the things I like around here.
A fine weekend for You.
Michael
Alfred: It is fun and pleasure browsing Your birding photos. I like Your idea of using the juxtapose presentation with the honey eaters sand plovers and tattlers.
Hi Ben, hi Alfred,
I see, You're being optimistic ;-)
No, it still doesn't work for me. I believe, what we see in phpinfo() is the apache's server environment which is different from the system environment. I will ask them to do me the same favour they did for Alfred.
But I can be convincing towards hosteurope. If neccessary I will just offer them to stay their customer.
I'll drop them a line tomorrow and will report after success.
Alfred, thank You for dealing with the ticket.
Michael
Hi Ben,
sounds interesting.
Is there a possibility to specify a tmpdir for sqlite from within backlight or should I beg Hosteurope to do me the same favour they did to Alfred?
Cheers Michael
Hi Ben,
what I don't really understand is that i can publish an album with the minimal database in place, but as soon as I restore the regular version I get the error again as soon as I clear the template cache. There must be a difference between the two versions that causes the regular one to produce an error the minimal one does not produce.
Hi,
I just got home and tried something:
I deleted backlight/data/designer/master.sq3, logged into backlight and made sure, that backlight created a new, minimal instance of master.sq3.
I could then clear the template cache, which was not possible with the original master.sq3 present.
I created a new test album with put in a photo and published successfully. Of course the LR plugin offered me only the default template.
I then put the original master.sq3 back into place.
Clearing the template cache resulted in the same error as before..
Isn't this a sign, that the database file might be corrupt?
If yes is there a way to test this? Or even mend it? Restore the database file from a backup?
My last successful publishing was in March. The farthest reaching Backup I can access through the Hosters Admin panel is from June 2nd. Two days before I first encountered the error. If the master.sq3 had been corrupted during the last publishing that went without error message that was in March :-(
Alfred: Moin, moin. Not knowing if this will be of any use, I would recommend You to get hold of a backup before the dates rotate out of reach.
Log into KIS, go to Produktverwaltung - Webhosting - Konfigurieren - Restore & Backup - Webspace Restore
Choose a suitable date (You may also be limited to dates after June 2nd) if You can remember the date of last publishing that seemed to have worked, the backup of that day might have been the one that caught Your master.sq3 in an OK state (or the date before - who knows) You can provide the path to Your /backlight/data/designer as source. If You choose "Restore-Typ tar.gz" that will result in a compressed file in the root of Your webspace once the backup is finished. You can than download and extract that and pull the master.sq3 from the package. "Restore-Typ direkt" will replace the current files with the version from the respective backup date which may not be what You want. The whole process will take a while.
You can only have directories restored, not single files. So the tar.gz type is what I recommend.
Do You have access from where You are? Edit: I just realize that You're not on a holiday but live in Australia. I like Your birding photos by the way.
By for now.
Michael
Yes indeed. I'm with Hosteurope too. I had two contacts with their telephone hotline. In the second contact their second level support was involved. Unfortunately they did not put me through directly but the agent switched lines between us.
They claimed everything to be OK on their side and the cause of the database being a frequency of database accesses to high for sqlite. They proposed switching to mysql and as I told them that this would be far out of scope they proposed to add a "sleep" command in appropriate places.
I precisely aked them to have a look at the file /backlight/data/designer/master.sql with the "fuser" command to find out which process keeps the database locked and possibly kill that process, which they refused. I tried that myself on the ssh console, but I have no permissions to execute "fuser". They asked me to open a ticket with support@hosteurope.de first. I had been very occupied last week and did not yet do that because I want to make certain that I will be reachable for requests and questions which was not given last week.
Alfred: I first noticed the problem on June 4th. If You log into backlight now and try to clear the template cache, You will probably get the same error message. If You download and read the recent logfiles in /backlight/data/designer/logs/ You will also find entries like
[11-Jun-2018 18:17:52] ERROR: Update query failed with error: SQLSTATE[HY000]: General error: 10 disk I/O errorsql:'UPDATE `engine_view` SET`name`=\'loop.php\',`type`=\'dynamic\',`path`=\'\',`rename`=\'\',`passthrough`=\'no\',`modified`=\'2018-06-11 18:17:51\'WHERE `engine_id`=\'99\' AND `view_id`=\'view.loop\''params:array () in PdoExtended.php on line 551
You will also find entries beginning like
[05-Jun-2018 00:00:32] ERROR: SQLSTATE[HY000]: General error: 5 database is locked|#
Just use the search string "error".
If You contact them and want to refer to me in case they recorded my calls in their support database: My hosting package is a shared hosting plan with the package number 1139230, so the virtual server is http://wp1139230.server-he.de/
If You mail me at ttg@michilge.de, I will tell You my Kundennummer. I will not find the time to open a ticket before Thursday.
Good luck so far
Michael