I registered my first domain (jcink.com) in 2002 with a little-known company at the time “ParsDot.” This was 13 years ago. Like many, I really had no idea what I was doing, but they were cheap and easily available. This company was actually just reselling domains from a much larger entity at the time known as eNom.com. Somewhere along the way, “ParsDot” folded and was ‘merged’ with eNom, and since then I have been an eNom customer for probably 12 years now.
Who is eNom, Inc? They are a domain name registrar hosting 14 million domains, founded in 1997. They’ve been around for a long time and are a major reseller and DNS provider. If you’re buying a domain from somewhere; chances are you might actually be registering it with eNom or at least using their domain name servers. For example, the extremely popular NameCheap uses them for their “DNS System V1”.
DNS is not something I’ve really had to think about in the 13 years I’ve been with eNom. They operate dns(1-5).name-services.com and that is what all of my domains have been configured to use for the longest time. For reliable basic DNS, it “just worked.” Until it didn’t.
Starting early November 2014, DNS being down or unreliable at eNom began. Each time they’ve had interruptions, I’ve tweeted about it, going as far back as November 9th. Sometimes, it would be a total outage and other times it would affect some ISPs who cached bad data being sent from eNom. Often I received complaints from CenturyLink users having this problem more often than others. I was not the only one; other eNom and NameCheap customers were tweeting up a storm.
— Jcink (@Jcink) December 9, 2014
Usually the response from eNom has been quite poor every time; either they’ve given a very vague response with no ETA. In one instance where they were offline for 3 hours during the afternoon, they revealed a reason for some of the downtime that didn’t jar well with me:
Looks like cause of one @eNom issue today was bad DNS code push. Pretty rare but when it does happen we roll back to restore service ASAP.
— Steve Banfield (@stevebanfield) November 20, 2014
Who, in their right mind hosting 14 million domains does a “code push” to their DNS servers in the middle of the day? Right here I started to lose my faith in eNom; this would be excusable if they were new but they have been a at this since 1997 and obviously are making a lot of money. Nonetheless the thought of having to move DNS was a real pain. I was hoping that this fumble would be the last and that they’d have patched the issues. January 2015 was the last recorded incident with them before a major one that occurred this month, so I gave them the benefit of the doubt after being with them for so long. And then comes March 9th — all hell has broken loose at enom and at 2AM I’m getting notifications of downtime. DNS for jcink.com and jcink.net are failing lookups, dns1.name-services.com (and all of its buddies) are totally offline and not responding to requests. I hit up twitter as usual and find nothing; enom is totally silent.
Their lack of response caused them to hit the press at around 7AM! The Register did a write up on it:
This goes on for even more hours until eNom finally says something with no ETA. With the mess they have created at this point, some people were not going to have DNS resolution due to caching for 24 hours or more. But I was already way ahead; after no response for 3 hours in the early AM, I decided I was tired of eNom and was just going to jump ship. Downtime is not a good reason to leave a provider, but poor customer service is. Had they been keeping users aware of what’s going on then I would have had a totally different perspective; but instead there was just nothing on their twitter or website regarding system status. Remember, these guys host 14 million domains. They know better than this, so I am less inclined to just give them a pass.
Hours later when they began coming back online, my domain was still saying “servfail.” Why?
When you change eNom DNS away from their nameservers, they do not temporarily keep your current records..
This matters a lot. It makes leaving eNom.com’s DNS very difficult because you have half of the internet querying the old servers and the rest querying the new. For the old ones that haven’t updated, it’s a total blackout until they do. There is a slight workaround to this; set an extremely high Time To Live on your primary DNS records really high before you leave their services. However, if anyone doesn’t already have your domain name result cached, you are going to be “down” for them. I have not figured out a way around this; and I don’t think there is one beyond “don’t use eNom DNS.”
I consider it really poor to outright delete DNS records immediately when someone switches. This caused me headaches for around 24-48 hours; it seems everyone’s DNS servers decide to re-cache records at different rates over that period of time. AT&T took the longest to get jcink.com’s records cached correctly; while Google and OpenDNS took the shortest. Watching a DNS server switch hour by the hour was an interesting experience and I certainly learned a lot, but I’d rather have not had to learn “the hard way.”
I don’t know what’s happened at eNom that after nearly a decade of service they are unable to keep their users online. DNS is serious, often a service provided by our registrar for free that we take for granted: but it needs to stay up. I will never host another domain with them again, and at the end of the year I won’t be renewing any either. It is more expensive than other domain providers, but I’ve always stayed because switching just wasn’t worth it.
Truthfully, my disappointment is less about the service downtime itself and more about the total and apparent complete lack of care to respond promptly especially with how many users they are serving. In the end, it is still my fault for not leaving the service sooner, but when you’ve been with a provider for a over a decade you want to give them the benefit of the doubt. I did based on past history, but it was the wrong call. Shame on you, eNom.