So i have no idea what defines what the plugin will use for a hostname when it starts backing up. But as i mentioned before I am using CloudFlare. So here’s how it works:
my domain, http://www.khurshed.com — Uses CloudFlare NameServers. and Points to their IP, they handle Web requests and request what they can not cache from my servers IP.
both, direct.khurshed.com and ssl.khurshed.com are direct DNS entries bypassing any CloudFlare infrastructure, those resolve to the servers true IP, no CDN. I also noticed both backups, and I have no idea why sometimes it decides to use one DNS name or the other, but they are the EXACT SAME server and exact same DB.
Hi zaheer,
Really close to this now. I think I have a potential fix we can try.
Can you go into Plugins -> Editor in your blog, and in the top right select Online Backup for WordPress and click the “Select” button.
Then underneath it after the page reloads you’ll see Plugin Files, click the link for wponlinebackup/include/transmission.php.
If you scroll to the bottom and work your way up you’ll eventually get to this:
/*private*/ function Request( $method, $api, $data, & $xml, $body, $in_backup )
{
$q = array();
$api = urlencode( $api );
// Create the query string
$data['blogurl'] = site_url( '/' );
Change the site_url(‘/’) to this:
/*private*/ function Request( $method, $api, $data, & $xml, $body, $in_backup )
{
$q = array();
$api = urlencode( $api );
// Create the query string
$data['blogurl'] = 'ssl.khurshed.com';
This will stop it changing and we can see if it fixes.
One other thing. Do you have a plugin that allows ssl.khursed.com to be used for SSL? Is it “WordPress HTTPS”? I am wondering if it is this that changes the URL mid backup. The workarounds I’m working on now will prevent this but I’d like to know the plugin so I can be sure I can test it.
Thanks,
Jason.
Did that, Got this :
The backup failed: The server returned the following error message: The blog URL of your WordPress installation does not appear to be valid. Please note that this plugin will only work on public websites, and not within an intranet or home network. If you believe your blog URL is correct, please contact support. Please consult the event log for more information.
then changed to : http://www.khurshed.com and got this:
The backup failed: Communication with the online vault failed.
24th Apr 2012 6.41.30 PM Error An online request succeeded but was malformed.
Please contact support if the issue persists, providing the complete event log for the activity. Diagnostic information follows
Also, there is no plugin that allows ssl.khurshed.com for https, its not secure unless https:// is specified , all it is , is a Cloudflare bypass.
Hi zaheer,
I think that fixed the random failures now. I’ll roll out that fix in 2.2.2 which is due for QA testing soon.
So we’re back to the corruptio I fixed right at the start. Can you let the backup run again, I’ve fixed it again and put in some more logging to see what’s causing it.
Hopefully we’ll be all good to go soon. Sorry for the issues and I really do appreciate your help in making the plugin better and more reliable.
Thanks,
Jason.
I guess it did not like all that tinkering.. After all that they all fail now. After removing and reinstalling the plugin still getting this :
24th Apr 2012 11.44.49 PM Error The backup failed: Communication with the online vault failed.
24th Apr 2012 11.44.49 PM Error An online request succeeded but was malformed.
Please contact support if the issue persists, providing the complete event log for the activity. Diagnostic information follows:
Failed at: /var/www/wordpress/wp-content/plugins/wponlinebackup/include/transmission.php(244)
XML parser: Completed entries: 1
XML log:
New xml() object instantiated.
fetch() called.
Raw XML:
<?xml version=”1.0″ encoding=”UTF-8″?>
<SynchroniseItems Final=”0″>
<Item>
<Bin>1</Bin>
<ID>1</ID>
<ParentID>0</ParentID>
<Type>1</Type>
<Name>OBFW_Database.sql</Name>
<Exists>1</Exists>
<FileSize>1376657</FileSize>
<ModTime>1335307520</ModTime>
<Path></Path>
</Item>
<Item>
<Bin>2</Bin>
<ID>1</ID>
<ParentID>0</ParentID>
<Type>0</Type>
<Name></Name>
<Exists>1</Exists>
<FileSize></FileSize>
<ModTime></ModTime>
<Path>a:0:{}</Path>
</Item>
</SynchroniseItems>
Well after 6 failures, 2 successful again. Im happy to help any way I can, If your feeling generous you can increase my backup to 200 so I can backup my whole site 😉 . Any idea on what causes the random inconsistency? I’ve left it on hourly again so you can hopefully see whats going on.
Hi zaheer,
Found what might be causing it.
Go here and download the Development Version.
http://wordpress.org/extend/plugins/wponlinebackup/download/
Just write down your encryption details etc and uninstall the plugin. Then go to Plugins -> Add New and click Upload and upload this development version.
Try it out and let me know how it goes.
Regards,
Jason
Hey There, Installed the dev version.
First Manual backup gave this error:
25th Apr 2012 1.08.20 PM Error The backup failed: The server attempted to retrieve the data, but received junk from your blog. This can happen if your blog is not accessible over the internet. Otherwise, you may have a third-party plugin installed that is changing the backup data as the server tries to receive it. Please contact support if this is the case so we may improve compatibility.
However 2nd manual backup worked without issue, And I am once again leaving on hourly to make sure all is well.
Hi Zaheer,
Many thanks.
Were any backups run before that first manual one? Can you check the Activity log. Or was that the first manual backup you ran on the dev version?
It’s just i’m seeing the following since you activated the dev version:
12.52.10 PM – Dev version logged in to Online Vault (Online Backup Settings page submitted)
12.52.41 PM – Backup started (1st)
1.02.37 PM – Backup started (2nd)
1.05.11 PM – Backup started (3rd)
1.08.19 PM – Backup finished (3rd) – Received junk
1.11.42 PM – Backup started (4th)
1.35.37 PM – Backup finished (4th) – Received successfully
Just wondering what those two unknowns are at 12.52.41 and 1.02.37.
Regards,
Jason.
Those “Unexpected stops” i have no idea. i did not initiate them. The only thing i can hypothesize is that when i set them to hourly. they may have tried to initiate themselves.
The first Manual i ran at 25th Apr 2012 1.08.20 PM according to the log.
That one failed with the error above, I have since run two additional manual backups successfully 🙂 . I am still leaving it on hourly , probably for the next 24 hours to ensure successful all the way through. Do you want to know if it fails again?
Hi zaheer,
Not sure what might cause those to initiate themselves but I’ve fixed a broken check that should’ve stopped the 2nd one starting and should’ve shown you the 1st one running and not let you start a 3rd.
I’ve also sorted the times so they’ll display in local time.
I’ll drop these into the dev version later on but I think your blog should back up fine now so you can just wait for the full release, and I’ll let you know when it is so you can update to it.
Thanks,
Jason.
Fantastic, Im glad i could be a (small) part of making this plugin better! , By allowing users to host their own keys, you do truly over a very secure solution! If you could let me know when the Dev version is up, ill install it anyway . I would like to check it out 😉
Last thing i thought I would let you know.. Using InnoDB mYSQL databases instead of myISAM works, but the activity log events can not be viewed. It seems all elements are OK in innoDB except wp_wponlinebackup_items even after reinstalling the plugin (ripping out all the tables) it likes to go back in as MyISAM. Just telling you as an FYI
Hi zaheer,
When you say the activity log events can not be viewed, what do you mean?
The items table is intended to be MyISAM. InnoDB lacks some features we need on that table so we force MyISAM. So that is fine.
Jason.
