Good day, Cole. > I spotted a single failure of GremlinExecutorTest#shouldCancelTimeoutOnSuccessfulScript. This test is fundamentally sensitive to execution times which gives it a chance for flakiness. I suspect this is a rare and random failure from this particular run.
Alas, no. It is a repeatable failure. I can give you rights to the repository so you can trigger it on demand. > It's not clear to me why this single test is failing to setup the server correctly when all other tests in this class are succeeding with the exact same setup. This definitely warrants more investigation if that test continues to be a blocker in certain environments Could you provide more reasoning about this? If those tests are not mandatory, why are they part of the mandatory run? Would it not be better to ignore them and unlock by option? On Fri, Jan 16, 2026 at 9:48 PM Cole Greer <[email protected]> wrote: > Hi Andrii, > > I took a look at the build output, the main issue I spotted was from a > small bug in ScriptEngineCacheTest which I've corrected in > https://github.com/apache/tinkerpop/commit/47baf67bd171039d8448dfbf2b3e24722bc36431 > . > > I spotted a single failure of > GremlinExecutorTest#shouldCancelTimeoutOnSuccessfulScript. This test is > fundamentally sensitive to execution times which gives it a chance for > flakiness. I suspect this is a rare and random failure from this particular > run. > > There is one additional failure I spotted in the ubuntu-24.04-arm - JDK 17 > (zulu) run which I have not been able to reproduce or explain so far. I > have extracted the following from the test logs: > > Error: Errors: > Error: > org.apache.tinkerpop.gremlin.console.GremlinGroovyshTest.shouldNotEvalToDetermineIncompleteLinesToSubmitForRemoteConsole > Error: Run 1: > GremlinGroovyshTest.setUp:41->super$2$setUp:-1->AbstractGremlinServerIntegrationTest.setUp:54 > » NullPointer Cannot invoke > "jdk.internal.platform.CgroupInfo.getMountPoint()" because "anyController" > is null > Error: Run 2: > GremlinGroovyshTest.tearDown:50->super$2$tearDown:-1->AbstractGremlinServerIntegrationTest.tearDown:61->AbstractGremlinServerIntegrationTest.stopServer:65 > » NullPointer Cannot invoke > "org.apache.tinkerpop.gremlin.server.GremlinServer.stop()" because > "this.server" is null > [INFO] > [INFO] > Error: Tests run: 41, Failures: 0, Errors: 1, Skipped: 0 > > It's not clear to me why this single test is failing to setup the server > correctly when all other tests in this class are succeeding with the exact > same setup. This definitely warrants more investigation if that test > continues to be a blocker in certain environments. > > I hope that this is helpful. > > Regards, > Cole > > On 2026/01/16 08:01:31 Andrii Lomakin via dev wrote: > > Good day, TinkerPop team. > > > > We have run TP tests on different platforms and distributions on JDK 17 > and > > found out that on some of them, TP tests fail. > > > > You can check results of the tests there > > https://github.com/JetBrains/ytdb-tinkerpop/actions/runs/21029624410 . > > > > > > -- > > Andrii Lomakin > > YouTrackDB development lead > > > -- Andrii Lomakin YouTrackDB development lead
