On Mon, 22 Jun 2015, Michael Grant wrote:

>    I'm running Debian Jessie (actually Testing).A  I have another machine
>    that's still running fail2ban 0.8.13.A  Here's a status from that machine
>    but it's not exactly the same usage so the difference could definitely be
>    that:

do you there also have that long list of jails and  *times?:

   # 90 days
   bantime A = 7776000
   # 10 days
   findtime = 864000

fail2ban needs to maintain a list of the IPs which it detected causing any
failed login... depending on how busy the server is (what fail2ban-client
status per each jail says?), in 10 days there could be quite a few.
Similar for bans but  I do not expect that many of them to cause a dent.

But indeed, machinery underneath 0.9.x became a bit more evolved (especially if
you use systemd etc)... at some point I was looking into analysis of what
consumes memory of a python process (using guppy IIRC).  It would be
interesting to see what consumes RAM in your case.  We already have an ability 
to
enhance logs with e.g. traceback "metric" for -regex and -testcases, e.g.

> bin/fail2ban-testcases --log-traceback -l info 2>&1 | head
Fail2ban 0.9.2.dev test suite. Python 2.7.10 (default, May 26 2015, 13:10:44) 
[GCC 4.9.2]. Please wait...
 2015-06-23 12:25:58,824 140640723736320 fail2ban-testcases:120>utils:191 I: 
Skipping systemd backend testing. Got exception 'No module named systemd'
testAction (fail2ban.tests.servertestcase.Transmitter) ...  2015-06-23 
12:25:58,825 140640723736320 ...6>servertestcase:152,66>server:79 Starting 
Fail2ban v0.9.2.dev
 2015-06-23 12:25:58,826 140640723736320 ...>server:112>asyncserver:133 
Fail2ban seems to be already running
 2015-06-23 12:25:58,826 140640723736320 ...>server:114 Could not start server: 
Server already running
 2015-06-23 12:25:58,826 140640723736320 ...>server:121 Exiting Fail2ban
 2015-06-23 12:25:58,827 140640723736320 
...>servertestcase:152,68>server:144>jails:70>jail:76 Creating new jail 
'TestJail1'
 2015-06-23 12:25:58,827 140640723736320 ...>jail:77,99,133 Jail 'TestJail1' 
uses pyinotify
 2015-06-23 12:25:58,834 140640723736320 
...>jail:77,99,134>filterpyinotify:69>filter:553,633 Set jail log file encoding 
to UTF-8
 2015-06-23 12:25:58,840 140640723736320 ...>jail:77,104 Initiated 'pyinotify' 
backend


I wondered if someone would be keen to contribute adding support of different
log metrics for the main process?  You could sneak/borrow tiny code for vmem
reporting from what we have in PyMVPA (although that one doesn't use standard
python logging, so base yours on how we implemented TraceBack Formatter)
https://github.com/yarikoptic/PyMVPA/blob/HEAD/mvpa2/base/verbosity.py#L434
so we could get smth like

$> MVPA_DEBUG=CLF MVPA_DEBUG_METRICS=vmem nosetests -s -v 
mvpa2/tests/test_clf.py | head -10 
/home/yoh/proj/pymvpa/pymvpa/mvpa2/base/verbosity.py:439: DeprecationWarning: 
get_memory_info() is deprecated; use memory_info() instead
  mi = __pymvpa_process__.get_memory_info()
[CLF] DBG{RSS/VMS: 107448/391992 kB}:             Cloning 
<libsvm.LinSVM(C=def)>#139891514649808
[CLF] DBG{RSS/VMS: 107812/392388 kB}:             Cloning 
<libsvm.LinSVM(C=def)>#139891514649808
[CLF] DBG{RSS/VMS: 107976/392524 kB}:             Cloning 
<libsvm.LinSVM(C=def)>#139891514649808
[CLF] DBG{RSS/VMS: 108336/392912 kB}:             Cloning 
<libsvm.LinSVM(C=def)>#139891514649808
[CLF] DBG{RSS/VMS: 108348/392912 kB}:             Cloning 
<libsvm.LinSVM(C=def)>#139891514649808
[CLF] DBG{RSS/VMS: 108428/392912 kB}:             Cloning 
<libsvm.LinSVM(C=def)>#139891514649808
[CLF] DBG{RSS/VMS: 108504/393044 kB}:             Cloning 
<libsvm.LinSVM(C=def)>#139891514649808


if we see when it grows we might identify why ;)

-- 
Yaroslav O. Halchenko, Ph.D.
http://neuro.debian.net http://www.pymvpa.org http://www.fail2ban.org
Research Scientist,            Psychological and Brain Sciences Dept.
Dartmouth College, 419 Moore Hall, Hinman Box 6207, Hanover, NH 03755
Phone: +1 (603) 646-9834                       Fax: +1 (603) 646-1419
WWW:   http://www.linkedin.com/in/yarik        

------------------------------------------------------------------------------
Monitor 25 network devices or servers for free with OpManager!
OpManager is web-based network management software that monitors 
network devices and physical & virtual servers, alerts via email & sms 
for fault. Monitor 25 devices for free with no restriction. Download now
http://ad.doubleclick.net/ddm/clk/292181274;119417398;o
_______________________________________________
Fail2ban-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/fail2ban-users

Reply via email to