> On June 26, 2014, 11:46 p.m., Mike Tutkowski wrote:
> > I first had a chance to run this patch through a sophisticated test tonight 
> > and noticed an issue with zone-wide primary storage that's based on the 
> > iSCSI protocol.
> > 
> > This patch leads to iSCSI storage being filtered out for ROOT volumes, 
> > which is a legitimate use case in my scenario.
> > 
> > As such, I had to return the filter method to the 
> > ZoneWideStoragePoolAllocator class.
> > 
> > I left in the other modification to the AbstractStoragePoolAllocator.
> 
> Yoshikazu Nojima wrote:
>     I thought the root volume filtering logic weird, but I didn't dig into 
> the problem.
>     Since this is a good opportunity, I would like to rethink this logic.
>     which storage requires the root volume filtering logic?
>     I understand that SolidFire's storage doesn't require it, but what about 
> other iSCSI storage?
>     Is there any reason that other iSCSI storage cannot handle a root volume? 
> Does anyone know?
>     Even if SolidFire's storage is the only iSCSI storage that can handle a 
> root volume, I think SolidFire's storage should be exempted from the 
> filtering logic not because storage is zone wide, but because storage is 
> SolidFire's.
>

I'm pretty sure any vendor's iSCSI storage can handle root volumes as it's just 
data placed on a disk. I don't know why that logic is there.

Any thoughts on how this code ended up in 4.4 and 4.4-forward (since the review 
request is just for master)? If that would have went out to the field, it would 
have rendered my plug-in completely useless.

Thanks!


- Mike


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/22717/#review46838
-----------------------------------------------------------


On June 18, 2014, 12:10 a.m., Yoshikazu Nojima wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/22717/
> -----------------------------------------------------------
> 
> (Updated June 18, 2014, 12:10 a.m.)
> 
> 
> Review request for cloudstack, Mike Tutkowski and Prachi Damle.
> 
> 
> Repository: cloudstack-git
> 
> 
> Description
> -------
> 
> refactor StoragePoolAllocator#filter logic to enable hypervisor type check, 
> storage type check for root volume and avoid list check, and to support IOPS 
> capacity control in a cluster wide storage pool and a local storage pool.
> 
> 
> Diffs
> -----
> 
>   
> engine/storage/src/org/apache/cloudstack/storage/allocator/AbstractStoragePoolAllocator.java
>  ddbb5a4 
>   
> engine/storage/src/org/apache/cloudstack/storage/allocator/ZoneWideStoragePoolAllocator.java
>  8fb9c8d 
> 
> Diff: https://reviews.apache.org/r/22717/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Yoshikazu Nojima
> 
>

Reply via email to