daviftorres commented on issue #12171:
URL: https://github.com/apache/cloudstack/issues/12171#issuecomment-3755261120

   Hey @vladimirpetrov, thanks for looking into this.
   
   When we initially deployed the Zones, we were using one storage appliance 
(I’ll call it “old”). Later, after acquiring larger storage appliances 
dedicated to CloudStack (I’ll call them “new”), we had to replace them in CS 
Management.
   
   Migrating Primary Storage is straightforward: we add the "new" storage to 
the zone, live-migrate the instances using “Migrate with storage”, and once 
nothing is left, CloudStack allows us to remove the "old" storage from the zone 
via the Web UI.
   
   However, this is not what happens with Secondary Storage. Even after 
removing all assets, CloudStack still blocks its removal, mainly because of 
System VM templates.
   
   For our regular templates, if we set the "old" storage to Read-Only and make 
sure it’s not in use, we can delete the template reference from the Zone and 
then copy it back from another Zone. This forces CloudStack to place it on the 
"new" storage (since there’s no way to choose the target Secondary Storage).
   
   What we’ve noticed is that CloudStack seems to prioritize Secondary Storage 
based on the order in which they were added, always preferring the first one. 
This makes the first Secondary Storage effectively impossible to remove, 
forcing us to delete it directly from the database.
   
   As you can imagine, this leaves orphaned assets that we later have to locate 
and clean up manually (see 
https://github.com/apache/cloudstack/issues/11641#issuecomment-3746236056
    for details about the internal tool we built to handle this).
   
   So, after that background, here’s the direct answer to your question (as 
best as I can remember):
   
   We discovered that some instances had ISOs attached and were left that way 
by mistake since deployment. When we later powered them down and tried to start 
them again, the ISO could not be found. On the UI, the message `Command failed 
due to Internal Server Error`.
   
   In that situation, there was no clear indication that the VM failed to start 
because the ISO file was missing. We would like either a more descriptive error 
message, or for CloudStack to automatically detach the missing ISO and continue 
starting the VM.
   
   Another case where auto-detach would make sense is for Compute Offerings 
with HA enabled, as this would improve the reliability of the HA feature.
   
   Please feel free to ask more questions, and I’ll try to respond faster this 
time.
   
   Best regards,


-- 
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