Re: NPE after Re-Deploying the Cartridges

2015-11-09 Thread Isuru Haththotuwa
Hi Imesh, As per the discussion in the thread, the final fix I did was removing the logic that is doing the validation is parallel and doing it in sequentially. On Tue, Nov 10, 2015 at 3:52 AM, Imesh Gunaratne wrote: > A great finding Isuru! I didn't get the fix you have proposed. Can you > ple

Re: NPE after Re-Deploying the Cartridges

2015-11-09 Thread Imesh Gunaratne
A great finding Isuru! I didn't get the fix you have proposed. Can you please explain it bit further? Thanks On Mon, Nov 9, 2015 at 11:32 AM, Isuru Haththotuwa wrote: > > > On Mon, Nov 9, 2015 at 12:17 PM, Akila Ravihansa Perera < > raviha...@wso2.com> wrote: > >> Nice catch Isuru! >> >> There

Re: NPE after Re-Deploying the Cartridges

2015-11-09 Thread Isuru Haththotuwa
On Mon, Nov 9, 2015 at 12:17 PM, Akila Ravihansa Perera wrote: > Nice catch Isuru! > > There is a JIRA already reported [1] for this issue. I've marked it as a > duplicate. > > Shall we get rid of this PartitionValidatorCallable class? I think it > adds unnecessary complexity to validation logic.

Re: NPE after Re-Deploying the Cartridges

2015-11-09 Thread Gayan Gunarathne
Great finding Isuru! On Mon, Nov 9, 2015 at 12:17 PM, Akila Ravihansa Perera wrote: > Nice catch Isuru! > > There is a JIRA already reported [1] for this issue. I've marked it as a > duplicate. > > Shall we get rid of this PartitionValidatorCallable class? I think it > adds unnecessary complexit

Re: NPE after Re-Deploying the Cartridges

2015-11-08 Thread Akila Ravihansa Perera
Nice catch Isuru! There is a JIRA already reported [1] for this issue. I've marked it as a duplicate. Shall we get rid of this PartitionValidatorCallable class? I think it adds unnecessary complexity to validation logic. [1] https://issues.apache.org/jira/browse/STRATOS-1595 Thanks. On Mon, No

NPE after Re-Deploying the Cartridges

2015-11-08 Thread Isuru Haththotuwa
Hi Devs, The error [1] is reproducible with 4.1.4 Stratos release. This can be reproduced by undeploying the application + cartridge, and then re-adding them, followed by deploying them from UI. This is occurring due to a race condition, when we are validating the partitions in parallel using Part