[Cadre-politics] first summer projects

[Dan MacNeil]

CONTROL PANEL 		(Manny)
SERVICE MONITORING 	(Matt & Rob)
LDAP 	      		(John)
DOWNTIME DATABASE 	(Kamala)
SPAM FILTERING 		(John & TBA)
CUSTOMER SURVEY 	(TBA)
LOWELL DEEDS / UTEC 	(dan)
BACKUP		 	(john)
SLICE			(john)
VIRTUAL SERVERS		(TBA)

CONTROL PANEL 		(Manny)
Right now, Manny is working on confirming that there is not an existing 
web based control panel that meets our needs. He's taking the existing 
short easy to understand entry [1] at wikipedia and fleshing it out.

	[1] http://en.wikipedia.org/wiki/Control_Panel_%28Web_Hosting%29

Apart from serving us and the users of wikipedia, this will look good in 
his portfolio. Next steps would be to start installing or to start writing.

SERVICE MONITORING 	(Matt & Rob)
Mon is a program that runs on one server and periodically checks to see 
if email,web,database,dns,ldap,etc are running, if they aren't it will 
take some action, send an email/text message/IM or trip a relay or whatever.

In their first day, Rob & Matt got Mon [2] installed on a workstation 
and monitoring another workstation.  Next steps are listing the hosts & 
services we wish to monitor, defining the alert and re-alert levels we 
need, installing it on a dedicated boxes and moving it offsite to 
downtown and CA. From past analysis of downtime, this should reduce the 
time it takes for us to respond to issues.

	[2] http://www.kernel.org/software/mon/

The time consuming part will be adding config files to CVS, which 
requires learning CVS and creating an installation checklist.

One thing to note is that out of the box Mon won't catch some failures 
in our email system. It just probes to see if stmtp the server is 
answering on the SMTP (simple mail transport protocol ) port. It doesn't 
have a way to know that the virus scanning queue behind the mail queue 
is stalled.

LDAP 	      		(John)
Our downtime actually isn't that big but a big chuck of our recent 
downtime is due to the LDAP server halting.  John's working on creating 
a backup server. Eventually we may distribute caching LDAP servers in 
the way we now distribute caching DNS servers.

DOWNTIME DATABASE 	(Kamala)

Kamala's working on the database to automate the info now at 
http://downtime.thecsl.org . Odds are good, Matt & Rob will write a 
custom alert to insert mon alerts into

SPAM FILTERING 		(unassigned but probably going John's way)
This is really starting to get out of hand. We need to update our filter 
rules and create a process for keeping them updated. This could be 
complicated or it could be simple. Since the CS dept's  anti-spam 
appliance with a $20,000 per year maintence contract is starting to fail 
as well, I'm guessing "complicated".

We also need to start doing SPF [3] for at least the listservs. SPF is a 
way to say that certain servers (lists.thecsl.org) are authorized to 
send email and others (boris.stick-it-you.ru ) are not.
	
	[3] http://openspf.org/

SPF only indirectly reduces our spam. We'll get just a little less 
bounce back from messages that have us forged as senders. However if we 
want our mail delivered, we need to start with SPF. AOL, comcast, 
verizon are starting to score non-SPF mail as "SPAM-LIKE"

The final anti-spam task is to make things moduler enough that we can 
just drop another identical spam filtering box in the lineup to handle 
increased load.

CUSTOMER SURVEY 	(next free person)
Some stuff like SPAM and downtime people have been pretty vocal on 
already. --We don't need a survey to know about this stuff.

However, many important people, (that's no lady, that's my wife) have 
said that we need to ask our customers what the want. These important 
people are right. Future projects should be strongly influanced by 
customer desire.

Josh Murry's (thank you) last draft is attached. I can imagine moving a 
big chunk of it to a format like:

You have 100 points to spend:

	____ 	service improvement 01
		(needs 10 pts for 50% success chance
		 100 pts for for 100% success chance)

	____ service improvement 02

	____ service improvement 03

...allocate these points among the options would be a good idea.

LOWELL DEEDS / UTEC 	(dan)

July is the end of the fiscal year for lowelldeeds.com, they don't like 
to rack up expenses that confuse the bean counters then. I'll spend that 
time on UTEC. (yeah, you've heard that before)

BACKUP		 	(john)

To reach the next plateau, (3 copies of 30 days worth of backups for 
everything, 1 copy 1 mile away), John & I pretty much just need to move 
the offsite-backup server offsite.

Other plateaus like quick bare metal restores and removable media to 
thwart evil insiders are for the future.

SLICE			(john)

I think there is more work to do with this, but to my shame I've not 
talked to John & Linda enough to know what.

VIRTUAL SERVERS		(TBA)

Virtual servers [] let you run several operating systems on one physical 
machine. As we got a $3,000 grant to buy a server to do this we should.

	[] http://www.xensource.com/solutions/

...if nothing else we get the quick bare metal restores for free.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: CSL Survey[2].doc
Type: application/msword
Size: 24576 bytes
Desc: not available
Url : http://lists.thecsl.org/pipermail/cadre-politics/attachments/20060601/0ec81dab/CSLSurvey2-0001.doc