Hi Tim! Yes, it looks as bug. Thank you for investigation! Feel free to contribute. ;)
On Fri, Dec 1, 2017 at 8:39 PM, Timay <[email protected]> wrote: > Hey all, > > We experienced an issue when trying to establish a semaphore after a single > instanced client node goes down hard (kill -9). Which afterwards we cannot > acquire a permit on the existing semaphore. However, if the client is > redundant, the permit is transferred successfully. > > I created a modified test of the SemaphoreFailoverSafeReleasePermitsTest, > which will close the initial semephore ignite instance then try and acquire > a permit and fail. > > SemaphoreFailoverNoWaitingAcquirerTest.java > <http://apache-ignite-users.70518.x6.nabble.com/file/t1401/ > SemaphoreFailoverNoWaitingAcquirerTest.java> > > I created a jira (https://issues.apache.org/jira/browse/IGNITE-7090) to > track as well, may try and dig further but wanted to get it out to the > group. > > Tim > > > > -- > Sent from: http://apache-ignite-users.70518.x6.nabble.com/ >
