OpenSSH exploit rumours swarm
As milw0rm shuts up shop
Rumours are circulating about the active exploitation of systems running older versions of OpenSSH, the open source remote administration utility.
Security watchers at the SANS Institute's Internet Storm Centre report circumstantial evidence of a mischief, including a log ostensibly showing an attack in progress, posted last Friday. In the absence of actual exploit code nothing can be confirmed.
One anonymous tipster told ISC that an exploit against older versions of OpenSSH might be presented as Black Hat, without providing any evidence. The absence of evidence has fuelled, rather than contained, chatter about the issue on internet security forums.
Speculation points to a possible problem that doesn't affect the latest version of OpenSSH, upgrading to which is a sensible idea irrespective of current gossip (examples here and here) among security geeks.
Red Hat Enterprise Linux ships with OpenSSH as a component and may therefore need upgrading, H Security adds.
ISC advises sys admins to upgrade to the latest version (5.2) of OpenSSH. The rumoured exploit is different from a confirmed denial of service attack posted on hacking site milw0rm, ISC notes.
In almost related news, milw0rm decided to call it a day on Wednesday, closing a chapter of one of the net's best known exploit portal/security websites. A brief note suggests the crew behind milw0rm are too busy with other projects to continue maintaining the site. ®
Not so hard
"Red Hat Enterprise Linux ships with OpenSSH as a component and may therefore need upgrading"
yum update openssh*
Then for good measure: service sshd restart
Ooh, so hard a monkey could even do it.
@Gordon Ross - Smartarse
RedHat ships with OpenSSH 4.3 with the patches backported in, as opposed to most other Linux distributions who now ship the latest release.
This doesn't look at all right. That log (the second one linked) doesn't have an RHEL5 kernel and doesn't have the RHEL5 apache. Other things don't look quite right either. Just googling for the kernel version -- 22.214.171.124-grsec-hostnoc-4.0.0-x86_64-libata -- throws up a lot of stuff about this supposed exploit.
I'm not buying this until there's better evidence than one oft-repeated log of dubious veracity.