MJZ down...

Portal Home > Knowledgebase > Articles Database > MJZ down...

Posted by mfive, 03-16-2005, 01:13 PM
Woke up this mornign to find my main site and all my client sites down. tried going to the mjzhosting.net main site and it is down too, whch concerns me because its on a seperate server. Hey, Matt if your out there, whats up?
Posted by shagbird, 03-16-2005, 01:15 PM
Anybody else afflicted?
Posted by Maxo, 03-16-2005, 01:20 PM
I think you have posted in the wrong section. Hopefully the moderators will move it.
Posted by shagbird, 03-16-2005, 01:35 PM
Feel free to move it, but my site is with MJZ and it's down as well.
Posted by Dactyl, 03-16-2005, 01:39 PM
perhaps you can give us more info about MJZ ?? a url ?
Posted by shagbird, 03-16-2005, 01:41 PM
Sorry, mjzhosting.net. The main page is down as well. It seems there's been a crash. Either that or the internet gods hate me... Eh, could be both.
Posted by Dactyl, 03-16-2005, 01:43 PM
yep its down allright
Posted by Sheps, 03-16-2005, 01:51 PM
Moved to Web Hosting... On a side note, what domains do you both host with mjzhosting?
Posted by shagbird, 03-16-2005, 01:51 PM
www.thewebasylum.com Don't ask...
Posted by mfive, 03-16-2005, 01:53 PM
main site is: hosttrek.net a few client sites are radiostarfleet.com sithcountychoppers.com among others
Posted by David, 03-16-2005, 01:54 PM
"It's dead Jim!" registrant-firstname: Matthew registrant-lastname: Zacharias registrant-organization: MJZ Web Hosting registrant-street1: 888 Dominion Street registrant-pcode: R3G 2N8 registrant-state: MB registrant-city: Winnipeg registrant-ccode: CA registrant-phone: +20.47834012 registrant-email: mjz@mjzhosting.net
Posted by UH-Matt, 03-16-2005, 02:08 PM
Another one bites the dust.
Posted by AHFBWEB, 03-16-2005, 02:26 PM
For $1.25 a month you expect uptime??????????????
Posted by IGobyTerry, 03-16-2005, 02:39 PM
Seems as if the server is sort of back up. MySQL appears to still be down however.
Posted by BigBison, 03-16-2005, 03:06 PM
...must...resist...urge...to...comment... Can't. http://forums.mjzhosting.net/viewtopic.php?id=255 http://forums.mjzhosting.net/viewtopic.php?id=260 So the server has trouble, but instead of working that out with the host it's time to cancel and transfer the accounts to a new server. This happens all the time at MJZ -- unless I'm mistaken, for some customers on this server this move will be what, the fourth or fifth since November? This time it's apparently all Sagonet's fault... I'm not surprised with the downtime reports here, these customers have been told in those threads to change their nameservers and IP addresses and we all know how smoothly that goes on less than 10 hours notice (that's how long the info on new nameservers and IP addresses has been available, but only if the forums are up). ...must...stop...now...
Posted by mfive, 03-16-2005, 03:08 PM
all is well in the universe: http://forums.mjzhosting.net/viewtopic.php?id=246 All in all im VERY happy with mjz, just the downtimes that have been happening lately have caused some friction (and not the good kind) with some of my customers.
Posted by BigBison, 03-16-2005, 03:16 PM
mfive, please consider that if your host had properly configured nameservers, a reboot wouldn't have taken out his entire operation like that. Having both nameservers on the same class C range, and having DNS, e-mail and webservers all on the .18 box, means you can look forward to reboot-related downtime every time that machine is patched. When that machine is down, domains would still resolve if the secondary nameserver were configured properly. Given the reliance on one nameserver, why make a change which will require a reboot while customers' changes due to the server move are propagating?
Posted by W3Solutions, 03-16-2005, 03:26 PM
Thats right , but after having downtime, its really NOT something to be happy about. Even your customers would be calling you. I lost a few of my clients myself bcoz of my server having problems, Once ive switched them to the new one, they refer other clients too. Im having exeptional uptime and giving good support.
Posted by AHFBWEB, 03-16-2005, 04:04 PM
I have noticed this as well, seems to happen about once every month or 6 weeks(How often are bills due?...just curious). If I am a reseller why would I want to buy from a reseller that for one reason or another cannot keep his own hosting accounts setup. Even worse, imagine being a customer of one of his resellers. downline nightmare
Posted by Odd Fact, 03-16-2005, 04:37 PM
Threads Merged.
Posted by coight, 03-16-2005, 08:45 PM
He is asking HIS customers if HE should cancel the server wtf? Looks like another host without any sys admin experience.
Posted by mmaaaaattt, 04-01-2005, 11:36 PM
Ahh, guess I'm a little late in catching this thread. Now that I've found it, I'll clear some things up for the record. The above comments about the troubles with the SagoNet box were completely unrelated to the mentioned downtime by the OP and other client in this thread. The downtime that they speak of in this thread occurred on 216.32.67.18, hosted in SAVVIS, the server they were both on. The Sago server move is an unrelated issue. As for the actual issue at hand, 216.32.67.18 had gone down due to a very CPU intensive website overloading the server. These things can happen from time to time in the hosting world, as everybody knows, hence AUP rules regarding fair resource usage. The site has since been removed and the server has been serene since. Once again, .18 was not related to the latest server move in any way. Last edited by mmaaaaattt; 04-01-2005 at 11:39 PM.
Posted by BigBison, 04-02-2005, 01:27 AM
You're saying none of those customers on the Sago box had their domains using the .18 box as nameserver? When you update your customers' IP addresses, their domains will take a while to reflect the change, depending on the TTL values of the various A records. While your affected customers were doing this, you rebooted the server serving as nameserver for some of them. To make matters worse, that .18 machine functions as both primary and secondary nameserver for your domain and others. When it's down, names don't resolve for customers on your other servers whose DNS records are hosted on .18. From RFC 2182: The Planet got burned by this same problem yesterday, by having their primary and secondary nameserver on the same subnet whose power failed. At least their nameservers are on different machines. Matt, you have machines at more than one datacenter, right? Designate a box other than .18 (.19 is the same box as .18, right?) as your secondary nameserver and learn how to set up zone transfers between them. Then, whatever your reason for tinkering with your primary nameserver while clients were updating their IP addresses on it, taking it offline won't disrupt DNS services for your customers, or your own email. I wish you would just admit that mucking with .18 should have waited a day, until those DNS updates settled down. On to e-mail. RFC 1912 recommends having a reverse DNS entry for your mail servers. You really ought to have a reverse DNS entry for that .18 machine, considering all it's doing for you. That other machine that's now your secondary DNS server now also needs to be your MX 20 system (the primary DNS system can remain MX 10, or whatever numbers you choose to designate priority) so that in the event of a failure of MX 10, e-mail for you and your clients hosted on your e-mail system won't lose e-mail. You can either synchronize the account IDs on both mail systems, or set up the MX 20 to store and forward, when MX 10 comes back online customers log onto it and get their mail normally, its delivery is just delayed. Either way, make sure you can send and receive e-mail from your client using MX 20. That way you can still communicate with your customers when .18 goes down. The problem with having your DNS go out like that, is any e-mail sent to you bounces, and that does not instill confidence in (even your) customers. You seem to be of the opinion that best practices are based on faulty reasoning by greedy corporate interests. I don't mind if you feel that way about some standards or standards bodies, but if that's what you think about the foundational RFC's which make the Internet function, then you're awfully ignorant of the RFC process. The best practices for operating nodes on the DNS system or a mailsite are common sense pragmatism, and describe exactly how to compete with the big boys on a shoestring budget. Look, you want to market yourself as a competent, low-budget hosting alternative. More power to you. But I can't take you seriously if your DNS and E-mail both go down when you reboot your server. That's called a single point of failure. There is no excuse for it, other than ignorance of the applicable foundational RFCs of the industry, for someone in your position with servers in multiple datacenters -- because it wouldn't cost you anything extra to do it right. http://www.dnsreport.com/tools/dnsre...mjzhosting.net But the OP says he isn't on .18: So, mfive's site, which isn't the Sago box or .18, was down. What mfive failed to mention in his post, and I should have asked him, was if he could access his account via its IP address. Was the server he's on down, or were his domains down because they were hosted on .18? Last edited by BigBison; 04-02-2005 at 01:38 AM.
Posted by mmaaaaattt, 04-03-2005, 12:10 AM
mfive lists his sites, they were always, and are still on .18. DNS: Looking up (hosttrek.net) DNS: Resolved hosttrek.net to 216.32.67.18 That's right, none of them were using the .18 box as nameservers. I think you're getting these incidents mixed up. I had Serverwizards apply the latest GR Security patch on .18 and when the box was rebooted, it didn't come back up right away - there were some issues with the kernel. I'm not trying to blame Serverwizards here, they have done excellent work for me in the past and I hold them in very high regard. Downtime will happen, mistakes will happen, it's a part of life and a part of the business. The Sago incident is different. The server was canceled due to a load of different problems including endless latency/downtime/DNS issues. The server would be down for a small portion of clients, yet up for others at the same time. Sites were not resolving in certain parts of the world - there seemed to be a lot of problems in the UK. i.e. people from all over the UK could not view a certain website on the server - it would appear offline, yet people from the USA/Canada could view it flawlessly. Weird routing bottlenecks and whatnot. The server's IP also apparently showed up in spam databases, which a client pointed out to me. He was not very pleased about this and subsequently requested his site to be moved to a different server. As for the latency issues, here is an email written to me from a client. His previous host had warned him about the server: This particular client was having horrendously slow loading website issues intermittently, as were many others. The server was brand new at the time and far from overloaded. The above problems are just some blatant examples. There were many more agitating little problems going on. Overall, it became far too bothersome and simply wasn't worth it, business-wise, which lead to the final decision of cancellation. I've gone way off on a tangent here, but I'm just trying to point out that the Sago issues you speak of are unrelated to the OP's downtime concerns in this thread. You're absolutely right about this part. It was a mistake for me to move my main hosting site to .18 in the first place. I'll be working to fix this...