Re: [VOTE] Release Apache Curator version 5.2.0

Not sure what happened there Enrico, I got an email saying the artefacts were pushed, but I couldn't see them either. It should be working now. cheers On Tue, Jul 20, 2021 at 7:57 PM Enrico Olivelli wrote: > Cameron, > thank you for running the release > > I cannot find https://dist.apache.org/

Re: [External Sender] Double Locking Issue

> In our case, it wouldn’t throw any exception because it had gone past > “creating lock nodes” and was blocked on wait(), which would only then be > interrupted when curator watcher notified on previous sequence node delete > event. So, you're using the version of acquire() without a timeout?

Re: [External Sender] Re: Double Locking Issue

Thanks Jordan for coming back on this. Please find my inline comments. Also provided below few additional version info that we are using, * Curator version – 2.5.0 * Zookeeper version – 3.5.3 > I don't see how this is the case. If the client has received a network > partition it shoul

Re: [VOTE] Release Apache Curator version 5.2.0

Cameron, thank you for running the release I cannot find https://dist.apache.org/repos/dist/dev/curator/5.2.0/ dist - Revision 48898: /dev/curator - .. - 5.1.0/ Il giorno mar 20 lug 2021 all

Re: Double Locking Issue

A few more things... > Based on our initial analysis and few test runs, we saw that Curator > acquire() method acquires the lock based on “about to be deleted lock node of > previous session”. Explanation : Ephemeral node created by previous session > was still seen by client that reconnected

Re: Double Locking Issue

A few things... SUSPENDED in Curator means only that the connect has been lost, not that the session has ended. LOST is the state that means the session has ended Be aware of how GCs can affect Curator. See the Tech Note here: https://cwiki.apache.org/confluence/display/CURATOR/TN10

Re: Double Locking Issue

(crossposting to dev@zookeeper) Hi ZooKeepers, can anyone take a look at this problem an user found while using Curator ? Thanks in advance Enrico Il giorno mar 20 lug 2021 alle ore 09:01 Cameron McKenzie < cammcken...@apache.org> ha scritto: > hey Viswa, > I'm by no means an expert on this ch

Re: Double Locking Issue

hey Viswa, I'm by no means an expert on this chunk of code, but I've done a bit of digging and it certainly seems that you've uncovered an issue. Ultimately the root cause of the issue is the weirdness in the way that ZK is handling ephemeral nodes. I'm not sure if this is intentional or a bug, bu