It was getting late as I posted on the first set of articles which I cleared out of my README file last night.
Here are the second set of articles I read:
"The Ranks of Service", by Douglas Hanna, published in Ping!(most recent edition)
Douglas provides an overview of the various levels of support that are considered standard by most hosting companies. You all know the terms, "…Level 1, 2 and 3…or if you are fancy, primary, secondary and tertiary!!"
Douglas is not very fancy in his delivery of this article (thank goodness!). He just delivers the "nuts and bolts" of the various levels and what each level generally consists of. He does show how various companies change the definition a bit to meet their needs/expectations.
He goes on further to touch on the importance on mentor-based training in these support roles.
He concludes with some notes on escalation procedures and describes the simplicity of some methods and the intense complexity of others. He provides some examples of different methodologies at work.
The article is supplemented with a "tips" column bar for companies looking to refine their escalation procedures. Here is the list:
- Have the team be aware of the different strengths and weaknesses of their fellow team members.
- Outline job descriptions and responsibilities
- Have administrators watch incoming tickets
- Escalate quickly
- Give updates
- Use training to your advantage
- Have a "go-to" person
- Share your statistics
- Monitor individual employees
"Server Backups, Are You Serious?", by Dave Young, published by Ping! (most recent edition)
Dave starts off by mentioning how, currently, most web hosting providers provide little, if any, backup of client data as part of their agreements with their clients.
The ones who don’t simply state that they view client data protection as the client's responsibility.
The ones who do often have hodge-podge mixtures of "FTP, tar, rsync and homegrown shell scripts".
He mentions how the word, "restore" doesn't mean that the client gets any data back, but rather that the client will get a new server wiped clean with a functional operating system.
He compares this against what other industries do for data backup. SANs, NASs, tape libraries, …vendors like EMC, Veritas, and Computer Associates. Doug has keen insights on why traditional licensing models of this software don't work for the hosting business.
He then discusses the data retention strategies for web hosting companies. Notably he discusses the impracticality of incrementals using TAR and Rsync.(too much disk space req'd).
Obviously retention plays big into disaster recovery scenarios and Dave mentions how some hosts have found out that their backups were failing, not to mention the failed server, and been in the unfortunate position of only having that single "backup". (oof!)
The remainder of the article highlights Righteous Software and its founder, David Wartell.
The history of the company and the creation of their technology Continuous Data Protection (CDP) is detailed and the efficiency of the CDP is lauded (hard disk sector based backups vs. file system level backups.
He mentioned how the RS group was able to efficiently store incremental backups using "disk safes"…aka "delta" technology.
So now, apparently, disaster recovery is narrowed down to four easy steps
- Fix hardware issues
- Boot the server into recovery mode
- Select the backup to restore
- Sit back and enjoy the complete disaster recovery.
Interesting read…seemed aimed more at the pure Linux-based networks, but I enjoyed the read!
I read another Q and A one pager on hosting questions…loved the Question and Answer on "blowing out the dust"!
The last one I read was a very brief interview with Open-Xchange Chairman Rafael Laguna. The one-page interview mentions that Open-Xchange has struck a deal with 1&1 and will have a low-cost open source option here.
Take care!