On one server, Fail2ban has banned over 150 IP addresses so far today
for attempting to hack into WordPress sites. (If you run WordPress and
you don't have protection against this, you need it!) The volume is such
that it managed to crash the MySQL server on one VPS today.
One suggestion I have seen is to use Apache's authorisation to limit
access to wp-login.php, but that a) involves telling a bunch of real
people a password and some of them could forget their own name and b)
increases the size of the apache process by about 20% over the slimmed
down version with as few modules enabled as possible that I have been
running.
Anyone have any other suggestions?
Ian
Totally OT for the list, and I'd be more than happy taking this off list at
the first opportunity.
Do any of you lovely folk have a Google Nexus 4 running stock Jelly Bean
4.2.2 (rooted)?
I changed over to a modified NFC stack (the detect NFC tag removal one) but
like a fool I forgot to keep a copy of NfcNci.odex, and I need both the
.apk and .odex to OTA to 4.3.
It's in /system/app and you'll need Root Explorer (or similar) to copy it
out to send it on.
Kind regards
Murray Crane
Hi all,
I was away on holiday for a while recently, during which time (on 21st
June to be precise) rkhunter started sending me daily report emails
like the one below, indicating that the perl and curl binaries on my
Debian 6.0.7 webserver changed. As far as I'm aware, my system only
gets updated when I manually perform it via apt-get, and I don't
remember doing that in the week or few preceeding the alert, so this
was a bit of a surprise. This report runs daily, yet the last update
I can see in /var/log/dpkg.log for perl is 2013-03-23, and 2013-05-10
for curl. It all seems slightly suspicious, and yet I have not found
any other evidence of the system being compromised. Network traffic
remains low, which I would expect to increase if it was hijacked.
Only thing I noticed is the OOM-killer kicking in a few times over the
last few months, possibly due to an Apache leak, but frankly I think
that's a bug somewhere rather than a symptom of a break-in, since it
started happening much earlier.
dpkg -s says that I have curl-7.21.0-2.1+squeeze3 and
perl-5.10.1-17squeeze6, and debsums says everything's OK.
# apt-cache policy curl
curl:
Installed: 7.21.0-2.1+squeeze3
Candidate: 7.21.0-2.1+squeeze4
Version table:
7.21.0-2.1+squeeze4 0
500 http://apt-cacher.lon.bitfolk.com/debian/security.debian.org/
squeeze/updates/main i386 Packages
*** 7.21.0-2.1+squeeze3 0
100 /var/lib/dpkg/status
7.21.0-2.1+squeeze2 0
500 http://apt-cacher.lon.bitfolk.com/debian/ftp.uk.debian.org/debian/
squeeze/main i386 Packages
# apt-cache policy perl
perl:
Installed: 5.10.1-17squeeze6
Candidate: 5.10.1-17squeeze6
Version table:
*** 5.10.1-17squeeze6 0
500 http://apt-cacher.lon.bitfolk.com/debian/security.debian.org/
squeeze/updates/main i386 Packages
100 /var/lib/dpkg/status
5.10.1-17squeeze5 0
500 http://apt-cacher.lon.bitfolk.com/debian/ftp.uk.debian.org/debian/
squeeze/main i386 Packages
The closest I can find via google is:
http://www.linuxquestions.org/questions/linux-security-4/rkhunter-warnings-…
but that doesn't seem to indicate a compromised system.
I just updated to rkhunter 1.3.8 from squeeze backports and it found a
few additional warnings, but all of them attributable to
non-suspicious causes.
Thoughts? I'm really loathe to re-install this system based on an
extremely vague suspicion.
Thanks!
Adam
---------- Forwarded message ----------
From: root <root(a)adamspiers.org>
Date: 20 July 2013 06:30
Subject: [rkhunter] coral.adamspiers.org - Daily report
To: root(a)adamspiers.org
Warning: The file properties have changed:
File: /usr/bin/curl
Current inode: 37410 Stored inode: 35028
Current file modification time: 1365866469 (13-Apr-2013 16:21:09)
Stored file modification time : 1333198916 (31-Mar-2012 14:01:56)
Warning: The file properties have changed:
File: /usr/bin/perl
Current hash: 400681f383f4a2b63d4615a8d7ad53<wbr>c2a685e3da
Stored hash : be5055e1642bec794804ebf8668a15<wbr>54864d218b
Current inode: 33794 Stored inode: 33812
Current file modification time: 1362591932 (06-Mar-2013 17:45:32)
Stored file modification time : 1361046751 (16-Feb-2013 20:32:31)
One or more warnings have been found while checking the system.
Please check the log file (/var/log/rkhunter.log)
Hi,
I've just noticed that my domains meowc.at and euroelection.co.uk,
both hosted on cosmo.bitfolk.com, don't seem to be working. (They were
working a few days ago.) I can ping them, but the websites are down
and I can't ssh to my VPS (meowcat) either.
I checked with http://www.downforeveryoneorjustme.com/ ansd not only
are my domains down, cosmo.bitfolk.com apparently is too.
What's wrong?
--
Phil Hunt, <cabalamat(a)gmail.com>
Hi everyone,
Please can you recommend a domain registrar that won't treat me like poo and that won't force me to use their name servers so I can host my own DNS? Reasonable pricing and someone that doesn't throw up needless obstacles to leaving would be a plus.
Thanks,
Paul.
--
Sent from my Android phone with K-9 Mail. Please excuse my brevity.
Some university famously has a stuffed toy in the corridor where the
computer science department has its offices. The idea is that you can't
bother the staff with a problem until you have explained it to the bear
- just explaining it to someone else often causes an 'ah ha' moment.
Drafting an email to this list has just been that bear :)
Today's tip: if someone complains their ad doesn't display on some site,
check to see if it's blocked by ad blocking browser plugins.
In this case, the name of the banner included '468x60' and this is
caught by some filters, because it's a common ad banner size.
Ian
Hi,
I am in transition from using a managed domain / web service for a couple
of sites.
I'll be cancelling the old web space account but not sure how to proceed
with the domain / dns hosting aspect.
Would appreciate any pointers from your experience.
Cheers,
Richard.
Hi everyone,
I'm a bit late to the Wheezy upgrade party, but successfully got the job
done over the weekend. This has been third dist-upgrade for this VPS,
having started life as an Etch box :-)
One minor gotcha for a VPS of this age is that it started out with a
xen kernel image (my /boot dir still has an Etch era
vmlinuz-2.6.18-6-xen-686 kernel), and Wheezy no longer provides a
separate xen kernel (in fact, it would seem that a xen kernel image
wasn't necessary in a domU in Squeeze).
Switching over to the default kernel from the linux-image-686 meta
package was therefore the way to go. This installs the 686-pae variant,
although I note that the Bitfolk Wiki page suggests the 686-bigmem
kernel, which has 4Gb+ memory support. Is there any particular benefit
to running a bigmem kernel on a smaller VPS with less than 4GB RAM?
--
Phil
On 06/06/2013 21:52, Ian wrote:
> I've got a Fail2Ban jail set up to ban anyone accessing any
> wp-login.php more than five times. It's just triggered a dozen times
> in a minute - there's another major burst of hack attempts going on.
>
> Especially if you or any clients have an account called 'admin' on a
> WP site - not a good idea, as it's the WP default and thus the
> primary
> one hackers go for - you want to watch out.
>
> Ian
>
> (Another eight triggers while writing this...)
I meant to post to the list about this too; I got hit on Tuesday to the
extent that my VPS OOMed.
After working out what was going on and adding to the fail2ban rules,
around 400 different IPs and around 2000 requests to wp-login.php were
blocked over the course of a couple of hours although it's died down
since.
If it helps anyone, my fail2ban filter:
[Definition]
failregex = [[]client <HOST>[]] WP login failed.*
[[]client <HOST>[]] client denied.*wp-login.php
The first line requires a change to your Wordpress theme to log failed
logins, described here:
http://blog.somsip.com/2012/02/using-fail2ban-to-protect-wordpress/
The second one comes from adding rules to .htaccess to deny requests
for wp-login.php and wp-admin to anything outside of the IP ranges I
use. The second rule should be sufficient; I added the first one a while
ago and didn't see any harm in leaving it.
Stuart