The Tech Bench: Controlling spam and mistakenly blacklisted IPs
Editor’s Note: For this week’s Tech Bench, we thought we’d share the particulars of a recent MST ticket. If you’ve ever struggled with administering email for clients over your VPS, this is one’s for you.
Recently, a customer submitted a ticket to the MST asking for help with the email on his VPS. It quickly came to the attention of our Director of Network Compliance, Mike Witty, when the customer explained that all email going through the mail delivery IP address on his server was being blocked by AOL, Gmail and AT&T.
The reason given by the email services for blocking his IP address was that it was a suspected source of spam. Yet the customer was in contact with all the users on his VPS and knew that they were not sending out spam.
The Tech Bench: The Ins and Outs of Nameservers
Nameservers are a small, but important part of your web service. They are essentially programs that translate your domain into an IP address and vice versa. They do this so that Internet users do not have to remember 32 bit IPv4 addresses such as 123.45.67.89 or 128 bit IPv6 addresses that are far longer. They instead allow us to recall and input the familiar .com domain names for all of our favorite websites.
Every domain name needs to be listed on a nameserver somewhere. There are three kinds of nameservers to choose from: nameservers run by your web host, third-party nameservers (typically, people use the nameservers run by their domain registrar if they have third-party nameservers), and private nameservers you run on your hosted server. But which to choose?
Read more
Customer Service in Hosting, Part 3: Customer Support From the Inside
As part of my role in Engineering, I’m the guy who ends up with all the next-to-impossible support requests. It may surprise you to learn that what makes these problems hard to fix often has little to do with the actual situation at hand. The biggest “technical challenge” I find is almost always insufficient data. And insufficient data usually stems from not being able to clearly understand from the customer what the problem is.
On behalf of customer support technicians everywhere, here are some tips to make sure that you always get the best possible help from your support team, in any company: Read more
The Tech Bench: File Permissions and Ownership
After last week’s discussion of PHP handlers, we thought we’d use this week’s Tech Bench to discuss the ins and outs of file permissions and ownership.
The Numbering Scheme
File permissions are often noted as a three digit number. This number is the binary shorthand used to classify the owner, group and public permissions for the file. The value farthest to the left is for the file owner, the middle value is for the group, and the value on the right is for anyone who does not fit into the owner or group and in most circumstances should be the most restricted.
The Tech Bench: All About PHP Handlers
What are PHP handlers?
In order for a site to run on PHP, the server must interpret the PHP code and generate a page when visitors access the website. It interprets the code based on which PHP library you are using, such as PHP 4 or PHP 5. A PHP handler is what actually loads the libraries so that they can be used for interpretation. PHP handlers determine how PHP is loaded on the server.
There are multiple handlers that can be used for loading PHP: CGI, DSO, suPHP, & FastCGI. Each handler delivers the libraries through different files and implementations. Each file and implementation impacts Apache’s performance, because it determines how Apache serves PHP.
It is essential for your server’s performance that you select the handler that fits your situation. Selecting the right handler is just as important as the PHP version itself. One handler is not necessarily better than another; it depends on your unique setup.
Note: In the event that your server runs multiple versions of PHP, you may assign different PHP handlers to each individual instance of PHP. For example, version 5 may be handled by CGI while PHP 4 is handled by DSO.
How to change the handler
Changing the handler on cPanel is very easy to do and takes only seconds. Log into WHM and navigate to: Main >> Service Configuration >> Configure PHP and SuExec
You simply select your PHP handler choice from the drop-down menu. Then hit “Save New Configuration”.
Note: If you do not see your desired choice in the drop-down menu, it may need to be compiled on the server first. You can do this via the EasyApache script of WHM.
List of PHP handlers
Customer Service in Hosting, Part 2: The Scope of “Support”
In the world of hosting customer support, “supported software and applications” is a phrase on the minds of many potential—and current—customers. Does my host support the software I am running on my server? Will they help me if it’s not working?
Truly supporting applications doesn’t simply mean supporting the software itself, though. It means managing the complex interaction between disparate software. This is way more important than any simple list of software that a web host can back away from in a pinch.
Let me give you an example: A lot of our customers use WordPress. At our suggestion, many of them install a caching plugin such as WP Super Cache or W3 Total Cache. Caching engines basically optimize server requests—decreasing the time it takes a page to load—on websites high in dynamic content by serving up flat data that looks dynamic. In our example, the client is running WordPress with WP Super Cache installed. So far, so good!
A few months after initial setup, the customer sees his traffic increasing, slowing down the server. He reads independently about the benefits of the PHP caching software APC and installs it himself. Not only does the site not get any faster, but it is now throwing up fatal errors left and right. So what happened?
In this case, the client didn’t realize that the default configuration for WP Super Cache does not interact with APC. He needed to configure both of these caching engines to work together. Also, months ago, he was worried about his PHP code being stolen so he installed Zend Optimizer, not knowing that APC and Zend Optimizer are completely incompatible. We’re now talking about the interaction of four pieces of software and how they work together and against one another on a server.
So: what’s ServInt’s responsibility here? Who “owns” the problem?
The Tech Bench: How can I troubleshoot “500 Internal Server Error” issues in cPanel/WHM?
500 Internal Server Errors can be caused by many issues. Please be advised if you are using Internet Explorer as your browser, it is known to display an internal server error when there isn’t one. Try using another browser first to eliminate this possibility.
If you are using a different browser, then the error may be due to one of the following: permissions/ownerships of files, memory usage, or coding (.htaccess, premature end of script, software specific).
To troubleshoot the 500 internal error you need to be ready to work on the command line via an SSH connection to your server. Read more
Customer Service in Hosting, Part 1: the Expectation of Service
Editor’s Note: With this post, our Director of Managed Services, Jim Tricarico, kicks off a three-part series on customer service in the hosting industry with a post about customer service from the hosting side. Check back in coming weeks for posts on the meaning of “supported” software and applications and the five things to always remember when submitting a support ticket.
Customer service is a tricky thing. Customers often judge the quality of care they receive based on how little they interact with a service representative. They often think that the quicker their problem is resolved, the better care they must be receiving—and, conversely, that they must be experiencing bad customer service when it takes more than a few hours to bring up a site that has gone down.
The truth may surprise you: often it’s the calls that take hours—or even days—that show you when your customer service team is giving you the best they possibly can. Let me explain, and in the process maybe shed a little light on how customer service in the hosting industry works.
To begin with, here’s a hard truth: sometimes our customers’ problems are so far outside the scope of our normal activities that there’s nothing we can do but say sorry and point them someplace else. For example, as odd as it may seem, more than once we’ve gotten requests from customers to help them troubleshoot the speed of their home Internet connection!
Pointing those customers in a different direction for their answers is easy because it’s clearly the best way to get them a quick solution. The gray area is when customers ask us to troubleshoot third-party software they’ve installed themselves, but can’t get adequate support for from the manufacturer. When these situations come up we try our best, but — well, maybe an example will help illustrate the difficulty of the situation.
I remember a customer whose site kept crashing. He was using a piece of software we didn’t officially support—and the sum total of the help he received from the software manufacturer was, “It’s not our fault, call your host.” We didn’t see the cause right away, but we agreed to work with the customer to troubleshoot the issue and spent a week peeling back one symptom at a time until the root cause was identified. In the end, this ticket took seven days to close out and had a total of 147 comments in it between our MST and the customer and from one tech to another. And yet, because this customer’s foremost concern was simply how long it was taking to fix the problem, he complained loudly about the customer support he had received.
Now, in this example we could have—and maybe we should have—spent more time explaining the causes of his site crashes and all the incremental steps we were taking to solve each one, but this is a delicate balancing act. Some customers want to know everything that is going on all the time, while others just want it fixed. Knowing how much a customer wants to hear when the problem is not yet solved can be tricky in the heat of the moment.
Make no mistake, there are incompetent and lazy hosts out there who run slipshod customer service operations. But you won’t find them by simply looking at how long it takes for them to “fix” your problem. If they’re continually trying new solutions and eliminating possible causes, they are doing exactly what they should be doing.
If you want to know if your hosts’ customer service is superior, ask yourself these questions:
- How many questions do they ask me when I have a problem? (The more the better.)
- Am I asking them to fix something that is completely out of their scope of support: unsupported software, issues with connectivity outside of the host’s network, etc? (Have realistic expectations about what your host can and will help you with.)
- Are they willing to at least look at my problem whether or not all of my applications are “supported software?” (A good host will acknowledge that there’s often no way to know the cause unless they’re willing to at least take a look.)
- If I ask, do they try to explain why something happened so that I can prevent it myself in the future? (A wise and valuable trait in superior customer service operations.)
- Have they ever simply told me they can’t figure it out and I’m on my own? (If they said this regarding “supported” software, it’s time to look for a new host.)
Photo by Jeremiah Ro
Hurricane Irene and Disaster Preparedness at ServInt
With Hurricane Irene projected to make landfall on the east coast of the United States in just a few days, many customers have been asking about what ServInt is doing to prepare for this potentially severe storm.
During any extraordinary event, ServInt operates under a set of Major Event procedures that have been crafted and refined over the last 16 years. These procedures include augmenting facility staff through overtime and multiple levels of on-call, assignment of task forces to deal with any contingencies that arise, and opening and maintaining lines of communication with our customers and vendors to ensure smooth operation throughout the event.
These procedures are important for making sure we’re always ready to provide our normal level of customer care in abnormal situations. But the bulk of our preparation for any given event takes place months to years earlier as we design and build our datacenters, choose our network providers and make our staffing decisions.
By choosing data center facilities that are expected to withstand any rationally conceivable event; by designing our network to quickly recover from the loss of any provider; and by making sure the staff we hire are reliable and steady in the face of the unexpected, we are able to work through most natural disasters without missing a beat.
In its 16 years, ServInt has weathered hurricanes, blizzards, tornadoes and—yes—even earthquakes. During Hurricane Isabel in 2003, while most businesses in the Washington DC metro area saw major disruptions, ServInt operated without customer interruption through the entire storm.
Frankly, our biggest concern during any natural disaster is making sure our employees can get to and from our facilities for their shifts. During any major event we roll out the cots and sleeping bags and stock the fridges. While we already run a 24/7 ship, we have to be prepared in the event that on-duty staff cannot make it home from a facility. Waiting for streets to be plowed or trees to be cleared is often our biggest concern.
Any web host’s true value is put to the test when extraordinary events occur, be it the explosion of traffic on a customer website, a hack or DDOS attack, or a natural disaster. These events and the host’s response are what customers remember. So as Hurricane Irene bears down on the eastern United States, rest assured that we are watching the weather carefully, unpacking our rain coats and dry-running all our disaster procedures. Our generators are fueled up and capable of powering your servers for a long time. Our cupboards are filled to bursting with cans of soup and stew. Our employees are ready. Your sites are as safe as any hosting company could possibly make them, and we’ll do everything we can to keep them up and running, no matter what mother nature throws at us.
Photo by NASA Goddard Photo and Video
Remembering the Macho Man
I was saddened to hear of Randall Mario Poffo’s untimely death earlier today. Of course, everybody knew Mr. Poffo as ”Macho Man” Randy Savage — or for me, just ”The Macho Man,” the professional wrestling character he made famous in the 70′s, 80′s, and 90′s. If you were a wrestling fan back then, it was tough not to root for Randy. He was brash, he was colorful, and he was the living embodiment of professional wrestling attitude.
I didn’t know him personally, of course, but in 2003 or thereabouts, when I was working in our Network Operations Center (pre-MST), the Macho Man was a client of ServInt’s. You can imagine the buzz around the office the day he joined the network: ”The real Macho Man?”, everyone would ask. ”Yeah, the real Macho Man.” I suppose to some there were “more serious” customers that one might use in a business or social setting to represent the quality and prestige of our company’s clientele. But for me, at that time, the first name that always came to mind — our coolest customer, in other words — was The Macho Man. And not surprisingly, people continued to ask, “the real Macho Man?” – and my response, delivered with a grin, would always be the same: ”yes, the real Macho Man.” There was no need for me to go any further. The name had been dropped. My mission was accomplished.
The real joy in having Mr. Poffo’s server under our wing was being able to talk to him on the phone on rare occasions. For us ’round-the-clock tech support guys, it was always a real treat when Mr. Poffo would call and ask us for help. You’d feel lucky if you were the guy who got to answer the phone when Randy was on the other end of the line. I remember him as a very curious and patient person, polite and professional. I guess I shouldn’t have been surprised that he was just a regular guy, but I was, if just a little. And if you’re curious… no, his every-day voice wasn’t the booming growl he used inside the ring, in TV commercials and back-stage interviews. But it was unmistakable — there was no doubt that you were talkin’ shop with Randy Savage.
After every Macho Man call, for a good 30 minutes or more, we couldn’t help but talk amongst ourselves in pathetic imitations of the Macho Man. It was impossible not to. Technical jargon just sounds better when you say it like Randy Savage. More fun, too. Every conversation was rehashed, sentence for sentence, for everyone else in the room. For some reason, it was incredibly important that each of us knew exactly why he called.
As much as I may have wanted to, I never had the courage — nor did I think it was appropriate — to ask him for a patented “Snap into a Slim Jim, oooh yeah”! But I was just as satisfied with phrases like “what version of Perl do I got” and ”let’s just reboot it.”
“Macho Man” Randy Savage was an icon of my childhood and my favorite client during my tech support years. On behalf of ServInt, I’d like to extend my sincerest condolences to Mr. Poffo’s family and friends. Thanks for all the memories, Macho Man.
Photo by goodrob13
