Here is an entry from my nightly analysis report:

New nsk.jdb failures (from 2007.08.22)
    nsk/regression/b4688375
        This test failed due to hang/timeout on Solaris SPARC-64 Server
        VM (machine vm-b1600s-4). It looks like the debuggee VM hung
        after continuing after the fourth breakpoint; no debuggee
        prompt after 5 minutes.

        Update: This test started at 2122 PDT and failed by 2133 PDT.
            /var/adm/messages on vm-b1600s-4 shows an NFS error from
            pkg.eng about 10 minutes before this test started and an
            NFS error from gtee/vmsge more than an hour after this test
            failed. This test appeared to use all local resources.

        Update: This test failed in the 2007.09.11 nightly on Win32
            Client VM (machine vm-amd64-04), but with a different
            failure mode. The test set the first non startup breakpoint
            and then continued, but no prompt was received.

        Update: This test failed while trying to write "cont" to jdb's
            stdin in the 2007.09.25 nightly on Solaris SPARC Server VM
            (machine nanonano). The following message is in the
            jdb.session log file:

                X connection to vmsqe.sfbay.sun.com:0.0 broken (explicit kill or server shutdown).

            I wasn't expecting there to be an Xsession with a jdb test.
            Win32 Server VM (machine vm-em64t-01) failed in a similar
            command sequence, but without the Xsession error mesg.

        Update: In the 2007.09.26 nightly on Win-AMD64 Server VM
            (machine vm-amd64-04), the test timed out after writing a
            'cont' command.

The above entry was with Dolphin-B18 JDK bits, but with a newer VM
so this is a Dolphin-B19ish failure.

Nightly search results for 2007.11 and 2007.12 only came up
with Win* timeout failures for this test.