Open hierony94 opened 8 years ago
Hello,
RioFS has a caching algorithm that tries to minimize access to the remote S3 cloud (so it increase the speed of filesystem operations). So we have to keep deleted objects in the cache untill we send the next update request to S3. If during that time you restore (or create a new) object with the same name on remote S3, then RioFS will get into "ambiguous" state: from one side your object is removed, but from other side it exists.
I'm aware about this issue, but currently this is one of RioFS limitations. My recommendations: try to avoid simultaneous delete and restore of objects. But if you have to: take a look at various "caching" timeout variables in the configuration file, try to play with these variables and see if it makes difference for you.
Hope it help!
Thanks for your response @wizzard ,
But It also happening without restore (or create a new) object with the same name.
I try this scenario (Without trying remove with rm -f
) :
@wizzard ,
Is it possible to disable caching mechanism in RioFS? I'm curious whether my problem still exist if I disable it.
I do it by change this line, I edit the value to false
<cache_enabled type="boolean">False</cache_enabled>
Is't the correct way to disable caching mechanism? Because after I try this, my problem still exist.
Please help me figure it out.
Hi @wizzard , I'm found a problem when deleting file from S3 Console
This is the scenario :
rm -f test1.txt
, No problem here.It show ?????????. Do you have any idea why it happen?
Thanks for your help.