There was a recent timeout of this test in window. See CODETOOLS-7900033. If you look at the timestamps in the test, it looks like it took just over 8 minutes, so the timeout failure is warranted. As for why it timed out, it seems to have run a bit slower than usual this time. Each iteration is taking about 30 seconds, which is a bit on the long side for this test on Windows. Looking at the mach5 history, I see other Windows test runs coming close to this timeout but finishing just in time. What is more interesting is the history of this test. It seems to be getting slower over time. Initially on 17 all runs were no more than just a bit over 1 minute, but crept up to 3 minutes during bug fix release testing, and the slowest runs are all in 2024. Runs for 18 are all 1 minute or less, but there has been no 18 testing since 2021. 19 spikes up to almost 5 minutes, with the slow runs coming Dec 2021 and Jan 2022 (no runs after that date). JDK 22 runs are as slow as about 3 minutes, but there have been no runs since 2022. JDK 21, 22, 23, and 24 all shows runs over 6 minutes (when including bug fix releases), with the slowest times all being in 2024. This seems to mostly be a Windows issue, although there is some indication of macosx times being slower for more recent runs.

This slowdown could be infra related (hosts having performance issues), SA related (something done to SA is slowing the test down, and those changes have been back ported to 17), or jtreg related (JDK-8268539 changed the test to run in driver mode in 2021, and was backported to 17). Possibly some combination could also be the cause.