When using a VM that is protected by the Fault Tolerance feature - what will happen if the first host fails at some extremly critical point of time and the secondary VM is just some milisecond after? I am thinking of some kind of database server or similar, what if something is sent over the network from a client and that action gets commited back to the client, and the microsecond later the host fails. When the secondary takes over - is there any risk for application error or corruption from this?
Hello.
Paraphrasing this or reducing it down to a simple answer won't do it justice. The answer lies in here.
Good Luck!
Hello.
Paraphrasing this or reducing it down to a simple answer won't do it justice. The answer lies in here.
Good Luck!
Thanks for the link. I have read it now and I do think I found the answer to my question. I think most of it can be found in the following text:
we have use it a lot and it works very very well even for those monitoring, batch transactional type of application. i would advise you testing it yourself as no one can really speak on behalf of your applications and needs.
idle-jam wrote:
we have use it a lot and it works very very well even for those monitoring, batch transactional type of application. i would advise you testing it yourself as no one can really speak on behalf of your applications and needs.
Thanks for the input. I am not thinking of any perticular application or service myself, just want to understand the general funcationality of FT.
Have you had any "real" failovers? :smileygrin: After reading the document (above) I realized that the techology is quite more advanced and the general problems that has to be taken care of is more than I thought. Impressive!
In short - No.
All writes are fully processed / committed at both ends - yet the CPUs run independently.
In other words, an application crash / bluescreen will kill both VMs, but the FT will not affect your application.