weizhouapache commented on PR #7405:
URL: https://github.com/apache/cloudstack/pull/7405#issuecomment-1804723604

   > > Agree with what Rohit said, this was enabled some releases ago, and it 
created issues for the Operators. It should be "impossible" by default, with 
the option to make it configured/enabled if the operator really needs it.
   > 
   > +1 what Andrija said. Not against having the feature, but let's make it 
depend on a global option or something, or a network offering like 
@borisstoyanov suggested.
   
   > We should the same feature that exists in a VPC as well to be honest.
   
   "remove" after "should"? 
   
   In my opinion,  redundant VR support is a very good feature,  it works 
pretty well,  if the environment has configured correctly and VRRP works 
without any issue.  
   
   However,  considering the various network devices,  multiple upstream 
routers,  VMware switch or openvswitch,  etc.  I have to say network 
configuration is too complicated ..... 
   
   I have the experience in the past that redundant VRs work well for some 
years but suddenly few (of many) networks have two MASTER VRs which lead to 
some downtime of user VMs.  The issue is not reproducible. After 
troubleshooting for some months,  the root cause was found, it was caused by a 
small configuration on one of the two core switches. 
   
   Be careful to enable redundant VRs in production... 


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to