mirror of https://github.com/fail2ban/fail2ban
- Updated
git-svn-id: https://fail2ban.svn.sourceforge.net/svnroot/fail2ban/trunk@362 a942ae1a-1317-0410-a47c-b1dcaea8d6050.x
parent
c80164b329
commit
f3a90b9021
|
@ -12,6 +12,8 @@ ver. 0.7.3 (2006/??/??) - beta
|
|||
- Added man pages. Thanks to Yaroslav Halchenko
|
||||
- Added wildcard support for "logpath"
|
||||
- Added Gamin (file and directory monitoring system) support
|
||||
- (Re)added "ignoreip" option
|
||||
- Added more concurrency protection
|
||||
|
||||
ver. 0.7.2 (2006/09/10) - beta
|
||||
----------
|
||||
|
|
63
TODO
63
TODO
|
@ -13,39 +13,25 @@ Legend:
|
|||
# partially done
|
||||
* done
|
||||
|
||||
- Verify TAI64N
|
||||
- Prefix all private functions with "__"
|
||||
|
||||
* Is there a do...while loop in Python? For interactive mode
|
||||
- Verify TAI64N
|
||||
|
||||
# implement all get/set functions
|
||||
|
||||
- correct handling of threads (join???)
|
||||
# correct handling of threads (join???), concurrency, etc
|
||||
protect "jails" in server.py
|
||||
|
||||
- signal handling (ctrl-c, etc)
|
||||
|
||||
* add a reload option to fail2ban-client
|
||||
|
||||
# see Feature Request Tracking System at SourceForge.net
|
||||
|
||||
* findall in dns.py should be no more needed
|
||||
|
||||
* remove utils/ directory
|
||||
|
||||
- improve installation process (better prefix support)
|
||||
|
||||
# improve documentation and website for user
|
||||
|
||||
* use Doxygen
|
||||
|
||||
- use PyLint to check the code
|
||||
|
||||
* better configuration files
|
||||
|
||||
* add a check to see if the time of the log messages is
|
||||
correctly detected (valid regexp)
|
||||
|
||||
* remove debug mode (root check)
|
||||
|
||||
# better return values in function
|
||||
|
||||
? use more email.Utils in mail.py
|
||||
|
@ -53,51 +39,10 @@ Legend:
|
|||
? add gettext support. Is this really needed for a server
|
||||
utility?
|
||||
|
||||
* send an email when fail2ban is running
|
||||
|
||||
* add multithreading. Python threading is not really
|
||||
efficient. However, fail2ban could benefit of it. We could
|
||||
use threads like this:
|
||||
- one thread which check for host to unban.
|
||||
- one thread per file to watch. This will allow things like
|
||||
different polling time for each file.
|
||||
<srv> is read-only (we only read log files) thus no locks
|
||||
are required. However, <meth> is read-write and must take
|
||||
care of concurrency in case of multithreading.
|
||||
|
||||
- add FAM/Gamin support. Should be quite efficient with
|
||||
threading. Take care that handle_one_event() release the
|
||||
Python lock.
|
||||
|
||||
# add a test framework. We could use unittest which is in
|
||||
Python since 2.1. It should be possible to run all tests
|
||||
automatically.
|
||||
|
||||
* add client/server using socket. Something similar to
|
||||
gdesklets. DBUS seems to be designed for desktop use.
|
||||
- fail2ban start -> start the daemon.
|
||||
- fail2ban stop -> stop the daemon.
|
||||
- fail2ban add <srv> <meth> -> add <srv> monitoring with
|
||||
<meth> ban method (iptables, hosts.deny, etc).
|
||||
- fail2ban del <srv> -> remove <srv> monitoring.
|
||||
- fail2ban status <srv> -> query current fail2ban status.
|
||||
Should return infos like a ban counter. Could be graph
|
||||
with rrdtool.
|
||||
- fail2ban pause <srv> -> suspend monitoring.
|
||||
- fail2ban resume <srv> -> resume monitoring.
|
||||
- fail2ban list -> list available services.
|
||||
- fail2ban flush <srv> -> flush the <srv> ban list.
|
||||
|
||||
* remove PID file.
|
||||
|
||||
* remove most of the command lines options if possible.
|
||||
|
||||
* add the possibility to specify wildcard in log files.
|
||||
Example: logfile = /var/log/apache2/access-*.log
|
||||
Should we start one thread per file or just one thread per
|
||||
serivce?
|
||||
We use just one thread per service
|
||||
|
||||
# autodetect date format in log file. Match the most popular
|
||||
format and sort them using the hit ratio. Should avoid
|
||||
user problem with regex and not have a big impact on perfs.
|
||||
|
|
Loading…
Reference in New Issue