Quote from: pyrographics at Dec 17, 2012, 07:03 PMRegarding my test file of semaphore locking, I put it on several sites without problem although one has still shown some cache errors. It probably won't work right if you turned off flock through the config since it ignores the existing non-flock workaround. It is 110% an experiment on my part so be careful with backing up first.
ok thanks. I did turn flock off on the config. I'll try without.
You can try this unofficial blow-up of xpdocachemanager.class.php out. Just change 8675309 in the sem_get functions to your own unique key. Does the key need to be unique for each file?
I now have it on all three sites I tried yesterday and they are operating without errors but that may just be because I broke flock not because I correctly implemented semaphor locking. Someone with more DEV experience than I will need to take a look at this.
Quote from: pyrographics at Dec 14, 2012, 11:14 AMYou can try this unofficial blow-up of xpdocachemanager.class.php out. Just change 8675309 in the sem_get functions to your own unique key. Does the key need to be unique for each file?
I now have it on all three sites I tried yesterday and they are operating without errors but that may just be because I broke flock not because I correctly implemented semaphor locking. Someone with more DEV experience than I will need to take a look at this.
2.2.7 didn't resolve any issues for me, but this file did. So far, any sites I have applied this to are no longer generating errors. I'm not going to pretend to understand why it's working, so I'm hoping someone can shed some light on this, or provide a more official fix.
☆ A M B ☆
One of my sites I applied the sem_get locking to did corrupt the web context cache this week so something is still wrong. No cache-related errors are displayed in the logs so I don't know what happened. Any updates from MODX/Rackspace on the issue?
MODX Staff
I've spoken to Chris at Rackspace. This is what he's said...
"We have provided what information we can to the developers of Modx about our environment in hopes that the devs of Modx can release a patch that would allow Modx to run on our environment. It's much easier for the devs to try and work around our environment than it is for us to make changes to our environment. We run multiple node web servers on a shared NFS system, and Modx uses some methods that are not particularly reliable in said environment.
From my research, it appears that the cache corruption issue has been prevalent for some time now and a recent iptables change on our side simply made it more "noticeable" and we're looking into alternatives for that issue on our side."
So hopefully modx devs will release a patch. Are there any modx devs that could inform us if you have all the information you need from Rackspace? And has there been any progress on the patch?
Many thanks,
Adam
MODX Staff
MODX Staff