Great Hack of 2008: Difference between revisions
mNo edit summary |
No edit summary |
||
| (4 intermediate revisions by 2 users not shown) | |||
| Line 15: | Line 15: | ||
== Timeline of Events == | == Timeline of Events == | ||
In an attempt to procrastinate from real work I've tried to document what happened on the week of no sleep, and insane amounts of coke. To give a bit of background, the hack happened just a few weeks after myself and johan were elected, lil_cain was the third admin, and werdz had recently moved from admin to webmaster. All of us were working full time on Intra, except for cian, who had a real job. Murphy was the main web server, running Ubuntu 6.06. Minerva was the main login server, also running Ubuntu. | In an attempt to procrastinate from real work I've tried to document what happened on the week of no sleep, and insane amounts of coke. | ||
To give a bit of background, the hack happened just a few weeks after myself and johan were elected, lil_cain was the third admin, and werdz had recently moved from admin to webmaster. All of us were working full time on Intra, except for cian, who had a real job. Murphy was the main web server, running Ubuntu 6.06. Minerva was the main login server, also running Ubuntu. | |||
Mostly, I've pieced this together from memory, email and irc logs. Some other people who were there may be able to fill in more of the details. | |||
--receive (May 2009) | --receive (May 2009) | ||
| Line 21: | Line 25: | ||
===Friday Morning === | ===Friday Morning === | ||
* Around 8.30am (ish) - People start to notice something is up. Most people are on the way to work and stuff. Pretty sure I got a text on the bus, can't remember who from. | * Around 8.30am (ish) - People start to notice something is up. Most people are on the way to work and stuff. Pretty sure I got a text on the bus, can't remember who from. | ||
| Line 46: | Line 49: | ||
* 10:30 - final mass hey goes out before shutdown | * 10:30 - final mass hey goes out before shutdown | ||
* 10:32 - access to the network is cut off. | * 10:32 - access to the network is cut off. A few root holders take refuge in #blueblock on irc.linux.ie to discuss meeting later on. | ||
=== Friday Afternoon === | === Friday Afternoon === | ||
| Line 69: | Line 73: | ||
Dano | Dano | ||
Secretary@redbrick.dcu.ie | Secretary@redbrick.dcu.ie | ||
=== Friday Evening === | === Friday Evening === | ||
The admins meet up in DCU about 5/6pm. Atlas also comes down. There is much drawing on a whiteboard about what our plans should be. Most of the night is spent taking backups. | The admins meet up in DCU about 5/6pm. Atlas also comes down. There is much drawing on a whiteboard about what our plans should be. Most of the night is spent taking backups. Getting some sort of mail processing in place is a priority, as we're not sure how long HEAnet/DCU will store our incoming mail for before giving up. Minerva's OS (and feck all else) is wiped and reinstalled by around midnight (after some extremely odd keyboard controller-related difficulties), and by the time we leave at around 1am, Deathray (not quite reinstalled, but isolated as much as possible from everything) is routing mail to inboxes. | ||
=== Saturday === | === Saturday === | ||
| Line 78: | Line 84: | ||
We meet up again in DCU early on Saturday and start the re-install process. One of the red dells is setup to allow admins external access to the network so stuff can be done remotely. | We meet up again in DCU early on Saturday and start the re-install process. One of the red dells is setup to allow admins external access to the network so stuff can be done remotely. | ||
Murphy continues to be shit. | |||
Cynic is re-installed with solaris, and DNS moves there. | Cynic is re-installed with solaris, and DNS moves there. | ||
| Line 133: | Line 139: | ||
===Monday === | ===Monday === | ||
Halenger builds ircd on minerva. Ubuntu starts working on carbon on Monday evening, after we discover that the ubuntu-server kernel hates the disk controller, and | Halenger builds ircd on minerva. Ubuntu starts working on carbon on Monday evening, after we discover that the ubuntu-server kernel hates the disk controller, and manually install a different kernel using a rescue shell. | ||
The new passwords went out on Monday. This required some epic hack scripting, since useradm was on one of the shite reddells, but that couldn't send email. | The new passwords went out on Monday. This required some epic hack scripting, since useradm was on one of the shite reddells, but that couldn't send email. | ||
| Line 143: | Line 149: | ||
I fell asleep on the bus back from work, and woke up at Ballymun shopping center. | I fell asleep on the bus back from work, and woke up at Ballymun shopping center. | ||
Murphy still refuses to work | Murphy still refuses to work. The Ubuntu installer CD refuses to talk to the upgraded version of the T2000 firmware (which we were told to upgrade by the installation instructions). | ||
{0} ok WORK DAMN YOU | {0} ok WORK DAMN YOU | ||
| Line 162: | Line 168: | ||
At this point, Deathray hadn't been touched yet, since mail needed to be migrated before that could happen, and Murphy was generally being a cunt, and wasn't working at all. | At this point, Deathray hadn't been touched yet, since mail needed to be migrated before that could happen, and Murphy was generally being a cunt, and wasn't working at all. | ||
=== Wednesday === | === Wednesday === | ||
| Line 168: | Line 175: | ||
Phaxx & atlas come to DCU to help make exim and mailman work. We also eat chinese. | Phaxx & atlas come to DCU to help make exim and mailman work. We also eat chinese. | ||
I (werdz) think it was around now (possibly Thursday or Friday) that I got ubuntu installed on murphy. Manually. After downgrading the firmware to version Poxy.ancient. The installer didn't seem to want to work at all (can't remember the exact list of problems), so it was installed using a rescue shell to create partitions, dpkg-bootstrap the system and install frivolous things like the kernel manually. Once this was done, Ubuntu booted, and things generally worked, but the serial console (well... what you got by typing console -f into the ALOM) randomly locked up. And some other things randomly broke. We decided this level of reliability wasn't what we were looking for in an OS. See [[Murphy]] for the rest of this epic tale, and how it eventually ended up with Solaris 10, a full year later. | |||
=== Thursday === | === Thursday === | ||
Svan gets fap working, and and syslog-ng doing remote logging. | Svan gets fap working, and and syslog-ng doing remote logging. | ||
=== Friday === | === Friday === | ||
Get mailman working about 1am. Continue testing mail and stuff until about 2/3am | Get mailman working about 1am. Continue testing mail and stuff until about 2/3am | ||
==== Saturday ==== | ==== Saturday ==== | ||
| Line 216: | Line 227: | ||
All the best, | All the best, | ||
andrew. | andrew. | ||
| Line 221: | Line 233: | ||
Welmar is backed up, and re-installed, followed by Obelisk. Anyterm starts working again late Sunday night. | Welmar is backed up, and re-installed, followed by Obelisk. Anyterm starts working again late Sunday night. | ||
=== Monday === | === Monday === | ||
| Line 227: | Line 240: | ||
We backed-up the rest of the stuff on Deathray, and re-installed that on Monday afternoon. Bitlbee was also setup late Monday night. | We backed-up the rest of the stuff on Deathray, and re-installed that on Monday afternoon. Bitlbee was also setup late Monday night. | ||
== The simplified version == | |||
[[Image:GreatHack.jpg]] | |||
[[Category:Redbrick History]] | |||
__NOTOC__ | |||
Latest revision as of 15:59, 26 April 2010
;_;
Redbrick died a death 25 April 2008, following the immortal words of lil_cain only a day before, on 24 April:
<lil_cain>: Thankfully in my time there have been no root exploits
Timeline of Events
In an attempt to procrastinate from real work I've tried to document what happened on the week of no sleep, and insane amounts of coke.
To give a bit of background, the hack happened just a few weeks after myself and johan were elected, lil_cain was the third admin, and werdz had recently moved from admin to webmaster. All of us were working full time on Intra, except for cian, who had a real job. Murphy was the main web server, running Ubuntu 6.06. Minerva was the main login server, also running Ubuntu.
Mostly, I've pieced this together from memory, email and irc logs. Some other people who were there may be able to fill in more of the details.
--receive (May 2009)
Friday Morning
- Around 8.30am (ish) - People start to notice something is up. Most people are on the way to work and stuff. Pretty sure I got a text on the bus, can't remember who from.
- 09:19 - people have posted to redbrick.computing.www about the problem.
- 09:27 - people in #lobby start to notice.
09:27 cambo anyone know what teh deal with the website is ? 09:27 marvin singer is afk 09:27 marvin tbh 09:28 fatwa cambo: haxed. lil_cain quit his admin position, that's his 'goodbye' revenge 09:28 tbolger entire site hax0rd ? 09:28 cambo fatwa: srsly... 09:28 fatwa tbolger: check www.redbrick.dcu.ie/~tbolger 09:28 fatwa cambo: ya rly 09:28 fatwa tbolger: it only seems to affect home directorys
- 09:31 - ssh access to murphy is shut down.
- 09:40 - apache is shutdown on murphy.
- 09:42 - access is traced to the hurling club account, which is disusered.
- 09:46 - checks on other servers reveal that auth.log is missing from minerva.
- 10:00 - the decision is made to shutdown all access
10:01 receive changed the topic of #lobby to: RedBrick access is about to be shut off, and will remain off until at least late tonight.
- 10:30 - final mass hey goes out before shutdown
- 10:32 - access to the network is cut off. A few root holders take refuge in #blueblock on irc.linux.ie to discuss meeting later on.
Friday Afternoon
D_fens, svan & drag0n go down to the server room and start tracing through the rest of the logs. The rest of us can't do much with no access.
Dano sends out a mail to the DCU RedBrick list
From: Damien Rathigan
Date: Fri, 25 Apr 2008 14:54:58 +0100
To: redbricksoc@list.dcu.ie
Subject: [Redbricksoc] downtime
hey all its me your beloved secretary,
There will be down time for unplanned maintenance of the network. We'll keep
you up to date with a dramatic blow by blow dialog to let you all know whats
happening! We're working on it and hopefully will not take too long! In the
meanwhile stretch out your legs get some coffee, go for a walk, smell a flower
or something,get high on life! And we'll be in touch!
Dano
Secretary@redbrick.dcu.ie
Friday Evening
The admins meet up in DCU about 5/6pm. Atlas also comes down. There is much drawing on a whiteboard about what our plans should be. Most of the night is spent taking backups. Getting some sort of mail processing in place is a priority, as we're not sure how long HEAnet/DCU will store our incoming mail for before giving up. Minerva's OS (and feck all else) is wiped and reinstalled by around midnight (after some extremely odd keyboard controller-related difficulties), and by the time we leave at around 1am, Deathray (not quite reinstalled, but isolated as much as possible from everything) is routing mail to inboxes.
Saturday
We meet up again in DCU early on Saturday and start the re-install process. One of the red dells is setup to allow admins external access to the network so stuff can be done remotely.
Murphy continues to be shit.
Cynic is re-installed with solaris, and DNS moves there.
Sunday
Minerva is finished off on Sunday. We also started scripting people's web directories back to normal. Temporary ldap is setup on one of the red Dells.
I attempt to send out an email to everyone with more information, but the web interface for mailman is down, and the thing refuses to let me authorise the mail without it. In the end I write a script to grab addresses from an ldif and just mail everyone.
From: root
Date: Mon, 28 Apr 2008 01:41:35+0100
Subject: [Webgroup] RedBrick Network Downtime
Hi,
By now most of you will be aware that we have been forced to take the entire
Redbrick network offline on Friday following a compromise in security.
There has been no loss of user data, email, websites, pr0n etc. - it is all
safely backed up. However, we are sure that the root access was gained on
both murphy and minerva as well ldap, so we are re-installing all systems
cleanly before they are going back online.
At present, the only service active is email - email which is autoforwarded
will be operating as normal, other email will be waiting when you login
again. Mailing lists are operational also, but the web interface is offline.
I am hopeful that we will be able to offer basic login services and chat
by Tuesday evening, but we can't make any guarantees - we are working as
fast as possible to get services back online safely. Progress with murphy
has been very slow, and website and database access will take longer to get
up again.
As ldap has been accessed by the intruders, all users will be issued a new
password in the near future. Please keep this password safe, you will
need it to log back in.
I would like to stress again that all your data and mail is safe and waiting
for you. If you need urgent file access or mail forwarding drop me a reply
and I'll do whatever I can.
We'll be in touch shortly when we have more online. We will also be keeping
the website error page on www.redbrick.dcu.ie updated with current info.
Thanks to everyone who sent us advice, encouragement, and steak. It is
much appreciated.
All the best,
andrew.
Monday
Halenger builds ircd on minerva. Ubuntu starts working on carbon on Monday evening, after we discover that the ubuntu-server kernel hates the disk controller, and manually install a different kernel using a rescue shell.
The new passwords went out on Monday. This required some epic hack scripting, since useradm was on one of the shite reddells, but that couldn't send email.
Tuesday
Phaxx gets IMAP working on minerva on Tuesday afternoon.
I fell asleep on the bus back from work, and woke up at Ballymun shopping center.
Murphy still refuses to work. The Ubuntu installer CD refuses to talk to the upgraded version of the T2000 firmware (which we were told to upgrade by the installation instructions).
{0} ok WORK DAMN YOU
WORK ?
At the end of Tuesday the network looked like:
Primary Login: minerva is ready Primary Services: carbon is installed. Needs work! Hey: Installed IRC: Done LDAP-secondary: Done WWW-temp: Done New ldap passwords: Done Reset changes to webtree: Done DNS-secondary: Working on cynic
At this point, Deathray hadn't been touched yet, since mail needed to be migrated before that could happen, and Murphy was generally being a cunt, and wasn't working at all.
Wednesday
I make some progress beating boards into working over the afternoon, it finally starts working properly about 7pm
Phaxx & atlas come to DCU to help make exim and mailman work. We also eat chinese.
I (werdz) think it was around now (possibly Thursday or Friday) that I got ubuntu installed on murphy. Manually. After downgrading the firmware to version Poxy.ancient. The installer didn't seem to want to work at all (can't remember the exact list of problems), so it was installed using a rescue shell to create partitions, dpkg-bootstrap the system and install frivolous things like the kernel manually. Once this was done, Ubuntu booted, and things generally worked, but the serial console (well... what you got by typing console -f into the ALOM) randomly locked up. And some other things randomly broke. We decided this level of reliability wasn't what we were looking for in an OS. See Murphy for the rest of this epic tale, and how it eventually ended up with Solaris 10, a full year later.
Thursday
Svan gets fap working, and and syslog-ng doing remote logging.
Friday
Get mailman working about 1am. Continue testing mail and stuff until about 2/3am
Saturday
Go down to dcu about 10am to put all the cables back to normal, and let users login again. Security finally let us in at about 12.30
SSH access to minerva is allowed again about 8pm, and the mail goes out shortly after.
From: Andrew Harford
Sent: Saturday, May 3, 2008 23:10:06 +0100
Subject: RedBrick Online
Hi,
We are pleased to announce that (parts of) the network are now back online :)
Minerva has been completly re-installed, and should be operating (relatively)
normally, with chat, boards and mail all working.
Carbon has also been re-installed, and it excepting logins, but there's
no configuration there, so don't expect it to do much. Mail, mailing
lists & boards are currently running from carbon.
Murphy was finally re-installed on Thursday, and we will begin work on
restoring web services and databases tomorrow, we hope to have them
operational shortly.
Severus & Obelisk haven't been re-installed yet, so at present there is
no nightly backups, or access to anyterm.redbrick.dcu.ie
Other services such as bitlbee, jakarta, ircproxy etc. will be coming
back after deathray has been re-installed. MUD is also offline.
All users should have received a new password to login, if you haven't please
contact admins@redbrick.dcu.ie
Despite our best efforts, we expect there to be many bugs & things missing.
If you find something wrong, please mail admins@redbrick.dcu.ie
All the best,
andrew.
Sunday
Welmar is backed up, and re-installed, followed by Obelisk. Anyterm starts working again late Sunday night.
Monday
Monday was a bank-holiday, so we got into DCU about midday.
We backed-up the rest of the stuff on Deathray, and re-installed that on Monday afternoon. Bitlbee was also setup late Monday night.
The simplified version
