(Happy New Year!)

It doesn't seem like very long ago when 4GB drives were considered HUGE.

With 1TB SATA drives becoming more popular now, and with lower price-points, more people seem to be thinking "hey, we have the space, why not use it". So rather than trying to be efficient with resources as they did in the past, people are not nearly as concerned with the size or number of files. I think it is great - bring on the mass storage at lower prices!

One challenge (and cost) though that many people don't consider is backing up that amount of space. With more and more things stored digitally, as a society we build a dependency on that data being available. So, consider a fairly simple redundant solution of RAID1. That could potentially give 1TB of redundant space - if a single drive fails, no problem, the data is still available. To protect (recover) though from an accidental overwrite, or from a total (very rare but possible) two-drive-failure, or to store some historical record of changing data - a backup solution is needed.

Daily backups with multi-week retention of just the simple, single-server, configuration of 1TB mentioned above can get complex. How fast does the backup run? How quickly can a restore be done? How much backup space is needed based on total space and retention requirements? Does the backup need to be stored off-site for disaster recovery? What is the cost of the drive(s)? What is the cost of the tapes? What is the cost of the backup software? What's the cost to manage it? What happens when there are two servers? How about two hundred servers, or two thousand servers?

Of course backups are just one factor of growing storage requirements. I plan on blogging about a number of the considerations in future posts.

These challenges can be addressed - and we do address them for our clients at ORCS Web. I'm just providing some "thinking points" for the people who wonder why managed redundant storage solutions are sometimes priced higher than they expect.

Happy Hosting!

~Brad

During our beta testing of our new VMWare environment, we created various volumes and then deleted them. After removing them from the Equallogic SAN, we found ESX server was still trying to login to those targets. After some research, I stumbled across the following files containing the stale entries:

[root@vmware root]# cd /var/lib/iscsi
[root@vmware iscsi]# ls -la
total 16
drwx------    2 root     root         4096 Dec 17 12:11 .
drwxr-xr-x   11 root     root         4096 Oct 23 09:17 ..
-rw-------    1 root     root          830 Dec 17 11:22 vmkbindings
-rw-------    1 root     root          474 Dec 17 12:16 vmkdiscovery
[root@vmware iscsi]#

Put the host into maintenance mode, then edit both the /var/lib/iscsi/vmkbindings and /var/lib/iscsi/vmkdiscovery files to remove the stale entries. Reboot the host and exit maintenance mode. Lastly, rescan the HBA:

esxcfg-swiscsi -s

Repeat for all hosts that are attempting to use the stale entries.

Out of a million or so emails received daily through our shared email servers, lately only about 6% is legitimate email. On holidays and weekends it is much worse (around 2% legitimate!). I don't know for sure, but I imagine those signs I occasionally see on the side of some roads - "Have computer? Work from home! Earn $$$!" - has a bunch of people kicking off spam scripts during their off-hours.

Does anyone actually read the spam? If so, does anyone really respond to it? I assume so or the spammers would have stopped bothering by now.

~Brad

A recent incident at CI Host reminds me that not all data center facilities are created equal. Because of that, here's a quick post with some information about the security at the data center that houses ORCS Web's managed servers:

Entering the building from the outside there is a station manned with a live person. Now inside the building, the person entering is still pretty far from any equipment. The next step is to enter what is sometimes called a "man trap". This area can only be entered by use of a proximity card and keying in a corresponding PIN number. Now in the "man trap" area the person is still not in the data center and has no access to any equipment. The next step is to pass another secure entrance that requires a proximity card and corresponding biometrics. Okay, now in the data center, all the equipment is still locked securely in cabinets and cages. Sounds a bit extreme to some, but we really can't be too safe when it comes to physical security. Hearing stories of situations at other data centers reminds us of that point.

Once inside, there are cameras all through the data center areas. Those cameras are actually monitored live onsite by multiple staff 24 hours per day - not just recorded as in some data centers. There are also vibration detection devices placed through the datacenter. I know these work because I've set them off myself. It's odd, in a good sort of way, to have a NOC person walk up to you to reprimand about setting off an alarm. I tripped an alarm and they quickly identified me on the cameras and came to tell me what I did. Very fast, professional, and cool.

Well, I just wanted to share some physical security information. Not everyone knows what levels we take to protect our customers.

~Brad

With 2007 winding down we are thinking more and more about 2008 lately. One of the things we are thinking about for 2008 is Microsoft Windows Server 2008. We have been running this next version of Windows in various contexts for most of the year - and we are very exciting about it!

The global launch is in just 84 days, so it will be here soon. If you are an administrator and haven't experienced it yet, you can download the release candidate version from Microsoft's site and go through a 30-day evaluation.

For those of you who have used it already - care to share some feedback about your experience on it?

Happy Hosting!

~Brad