[LUCID] We need "limit" Upstart-ed ssh respawning
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
openssh (Ubuntu) |
Fix Released
|
Undecided
|
Unassigned |
Bug Description
ssh is now handling by upstart native jobs(with /etc/init/ssh.conf configs), and "respawn" stanza. This is great works. But this behavior is not familiar for security paranoid.
Scenario:
- sshd is enougth to berief, but this is not perfect. In future, if we have got vuln for exploiting sshd, that can "randomly" type atacks (e.g.: buffer overflow, it is exploitable with probabilistic atacks, like brute force.).
- endless respawning is weaken for these "memory address brute force" atacks.
IMHO, when using "respawn" stanza, it needs something "limit" stanza (e.g.: respawn limit 60 30)
or any other way ( e.g.: MAC by AppArmor).
limitation by "limit" can mitigates(atack speed limitation) probabilistic atacks.
This limitation provides potential over-limit DoS, but unlimited respawning is dangerous.
# May be, we have to improvement upstart respawn stanza, like "respawn delay"
# feature...(see https:/
....Yes, this wish is not only ssh services. But ssh is most popular + internet accessible services.
In general cases, administrator use "ufw limit ssh" settings. Its not hazardous.
This bug was fixed in the package openssh - 1:5.3p1-3ubuntu2
---------------
openssh (1:5.3p1-3ubuntu2) lucid; urgency=low
* Always set child processes' OOM adjustment to 0, since Upstart will have
set sshd's OOM adjustment on startup and so simply restoring the startup
value won't work (LP: #293000). Thanks to Karsten Suehring for
analysis.
* Limit automatic respawning to 10 times in 5 seconds (thanks to Fumihito
YOSHIDA for the suggestion; LP: #533352).
-- Colin Watson <email address hidden> Sat, 06 Mar 2010 22:00:19 +0000