HI
I am running vdbench 5.0.4 on Contos 0.01 and I am seeing multiple failures with "Too many thread blocks". It is asking if I have too many threads for for files. I currently have 32 read and 32 write threads and 500 files in a single directory, here is the info that is pumped out, any help would be appreciated. These appears to have occurred when I added direct flag on open.
11:21:02.729
11:21:02.729 Message from slave localhost-0:
11:21:02.729 Thread: FwgThread write /mnt/test/vdbench rd=rd1 For loops: None
11:21:02.729
11:21:02.729 last_ok_request: Tue Aug 12 11:20:51 MDT 2014
11:21:02.729 Duration: 11.34 seconds
11:21:02.729 consecutive_blocks: 10001
11:21:02.729 last_block: FILE_BUSY File busy
11:21:02.729 operation: write
11:21:02.729
11:21:02.729 Do you maybe have more threads running than that you have
11:21:02.729 files and therefore some threads ultimately give up after 10000 tries?
11:21:02.734
11:21:02.734 **********************************************************
11:21:02.734 Slave localhost-0 aborting: Too many thread blocks
11:21:02.734 **********************************************************
11:21:02.734
11:21:04.510
11:21:04.510 Slave localhost-0 prematurely terminated.
11:21:04.510
11:21:04.510 Slave aborted. Abort message received:
11:21:04.510 Too many thread blocks
11:21:04.510
11:21:04.510 Look at file localhost-0.stdout.html for more information.
11:21:04.510
11:21:04.510 Slave localhost-0 prematurely terminated.
11:21:04.510
11:20:51.152
11:20:51.259 11:20:51.259 Beginning of run setup
11:20:51.259 11:20:51.259 **********************
11:20:51.260
11:20:51.260
11:20:51.271 11:20:51.271 Reading control file for anchor=/mnt/test/vdbench
11:20:51.272 11:20:51.272 Completed reading control file.
11:20:51.272 11:20:51.272 Completed the creation of the directory list for /mnt/test/vdbench: 1 directories.
11:20:51.272 11:20:51.272 Memory total Java heap: 123.500 MB; Free: 103.956 MB; Used: 19.544 MB;
11:20:51.300 11:20:51.300 checkMemory()
11:20:51.300 11:20:51.300 Memory total Java heap: 123.500 MB; Free: 108.714 MB; Used: 14.786 MB;
11:20:51.313 11:20:51.313 Generated 50 file names; total anchor size: 6.250g
11:20:51.313 11:20:51.313 Writing control file for anchor=/mnt/test/vdbench at start of run.
11:20:51.329 11:20:51.329 Completed control file for anchor=/mnt/test/vdbench at start of run. dirs: 1/1 files: 50/50 sizes: 6.250g/6.250g/0/0.0
11:20:51.330 11:20:51.329 During anchor creation for anchor=/mnt/test/vdbench there were 1 directories and 50 files
11:20:51.330 11:20:51.330 Reusing existing FileAnchor structure.
11:20:51.330 11:20:51.330 Skew for fwd=fwd1,fsd=fsd1,operation=read: 50.0
11:20:51.330 11:20:51.330 Skew for fwd=fwd2,fsd=fsd1,operation=write: 50.0
11:20:51.332 11:20:51.332 Started 32 threads for fwd=fwd1,fsd=fsd1,operation=read
11:20:51.337 11:20:51.337 Started 32 threads for fwd=fwd2,fsd=fsd1,operation=write
11:20:51.337 11:20:51.337 Started 64 FwgThreads
11:20:51.337 11:20:51.337 Waiting for task synchronization
11:20:51.337 11:20:51.337 task_wait_start_complete() end
11:20:52.000 Starting RD=rd1; elapsed=360; fwdrate=max. For loops: None
11:20:52.040 11:20:52.040 task_run_all(): 65 tasks
11:21:02.677 11:21:02.676 trace: 11:21:02.657 FILE_BUSY File busy
11:21:02.677 11:21:02.676 trace: 11:21:02.657 FILE_BUSY File busy
11:21:02.677 11:21:02.677 trace: 11:21:02.657 FILE_BUSY File busy
11:21:02.677 11:21:02.677 trace: 11:21:02.657 FILE_BUSY File busy
ditto
11:21:02.687 11:21:02.687 trace: 11:21:02.676 FILE_BUSY File busy
11:21:02.687 11:21:02.687 trace: 11:21:02.676 FILE_BUSY File busy
11:21:02.687 11:21:02.687 trace: 11:21:02.676 FILE_BUSY File busy
11:21:02.687 11:21:02.687 trace: 11:21:02.676 FILE_BUSY File busy
11:21:02.687 11:21:02.687
11:21:02.687 11:21:02.687 Miscellaneous statistics:
11:21:02.688 11:21:02.687 READ_OPENS Files opened for read activity: 32
11:21:02.688 11:21:02.688 WRITE_OPENS Files opened for write activity: 18
11:21:02.688 11:21:02.688 FILE_BUSY File busy: 140043
11:21:02.688 11:21:02.688
11:21:02.689 11:21:02.688 sendMessageToConsole: Thread: FwgThread write /mnt/test/vdbench rd=rd1 For loops: None
11:21:02.689 11:21:02.688 sendMessageToConsole:
11:21:02.689 11:21:02.688 sendMessageToConsole: last_ok_request: Tue Aug 12 11:20:51 MDT 2014
11:21:02.689 11:21:02.688 sendMessageToConsole: Duration: 11.34 seconds
11:21:02.689 11:21:02.688 sendMessageToConsole: consecutive_blocks: 10001
11:21:02.689 11:21:02.688 sendMessageToConsole: last_block: FILE_BUSY File busy
11:21:02.689 11:21:02.688 sendMessageToConsole: operation: write
11:21:02.690 11:21:02.689 sendMessageToConsole:
11:21:02.690 11:21:02.689 sendMessageToConsole: Do you maybe have more threads running than that you have
11:21:02.690 11:21:02.689 sendMessageToConsole: files and therefore some threads ultimately give up after 10000 tries?
11:21:02.690 java.lang.RuntimeException: Too many thread blocks
11:21:02.690 at Vdb.common.failure(common.java:295)
11:21:02.690 at Vdb.FwgThread.block(FwgThread.java:846)
11:21:02.690 at Vdb.FwgThread.block(FwgThread.java:762)
11:21:02.691 at Vdb.FwgThread.findFileToWrite(FwgThread.java:422)
11:21:02.691 at Vdb.OpWrite.doOperation(OpWrite.java:35)
11:21:02.691 at Vdb.FwgThread.run(FwgThread.java:157)
11:21:02.691 11:21:02.690
11:21:02.691 11:21:02.690 Too many thread blocks
11:21:02.691 11:21:02.690