[HTCondor-users] Interpreting XferStatsLog and understanding message flow

[George Papadimitriou <georgpap@xxxxxxx>]

Hello,

I'm trying to read and understand the log entries of the XferStatsLog, that present tcp connection statistics from condor io.

However, I'm a bit confused with the message flow and it's not entirely clear what does each log entry means.

Here is an example output from the log (excluding the tcp related fields), where I execute a workflow with 4 jobs (2 running at the same time, each time) and condor stages in and stages out files to and from the worker nodes.

Timestamp	Description (???)	JobId	files	bytes	seconds	dest
02/08/18 18:52:31	(1822.0) (523676): (peer stats from starter): 02/08/18 18:52:31 (1823.0) (523677): File Transfer Upload	1823	8	13070752	0	10.103.0.11
02/08/18 18:52:31	(1823.0) (523677): (peer stats from starter): 02/08/18 18:55:54 (1822.0) (523676): File Transfer Download	1822	5	196890	0.01	10.103.0.11
02/08/18 18:55:54	(1822.0) (523676): (peer stats from starter): File Transfer Upload	1822	5	196890	0.03	10.103.0.12
02/08/18 18:56:07	(1823.0) (523677): File Transfer Download	1823	5	196900	0.01	10.103.0.11
02/08/18 18:56:07	(1823.0) (523677): (peer stats from starter): File Transfer Upload	1823	5	196900	0.03	10.103.0.12
02/08/18 18:56:07	(1824.0) (524617): File Transfer Upload	1823	10	13248090	0	10.103.0.11
02/08/18 18:56:07	(1824.0) (524617): (peer stats from starter): 02/08/18 18:56:17 (1826.0) (524676): File Transfer Upload	1826	10	13248093	0	10.103.0.11
02/08/18 18:56:17	(1826.0) (524676): (peer stats from starter): 02/08/18 19:22:15 (1824.0) (524617): File Transfer Download	1824	3	1842522	0.01	10.103.0.11
02/08/18 19:22:15	(1824.0) (524617): (peer stats from starter): File Transfer Upload	1824	3	1842522	0.03	10.103.0.12
02/08/18 19:22:30	(1826.0) (524676): File Transfer Download	1826	3	1842569	0.01	10.103.0.11
02/08/18 19:22:30	(1826.0) (524676): (peer stats from starter): File Transfer Upload	1826	3	1842569	0.03	10.103.0.12

In this example the field I have marked as "Description" is the most confusing to me, especially because some records (eg. lines 1, 2, 7, 8) seem to have information about multiple jobs and I cannot understand the meaning.

Could someone explain the meaning of each record and maybe give me a high level description of the message flow ?

Thanks,

George