Notes about the best hosting: part three, problems and their solutions

With this note, I continue the cycle of articles on hosting with the host with the strange Mochahost name for the Russian ear. I recommend reading those who missed the first two articles:

• Best Hosting Notes: Part One
• Notes on better hosting: part two, helpdesk and setting up wildcard subdomains in IIS

So, if the first two articles had a generally positive attitude, then this note will consist of listing the difficulties and the troubles received.

Below is information on the problems that I faced for a month and a half of hosting. In the end I will draw some conclusions, but for now let's proceed to the listing of problems.

Problem 1. Mail

The first and most unpleasant moment in the work of the hosting was the failure of the mail at the very moment when I began to test my progg.ru service among my friends in Habr. Mail just stupidly stopped coming. No return letters were returned, so it was concluded that the problem was exclusively on the side of the hoster.

Solution of a problem. The solution of the problem took longer than I expected, about a day. During this time, I corresponded with the caliper, which in fact did not take much power. A request to solve the problem was sent, an answer was received that the request was accepted. After that, after a while, came another answer, saying that "check, we have excluded you from the" gray "list." Unfortunately it did not help, as I immediately reported. The third message from the caliper already accompanied the solution to the problem. Judging by the hoster's words, they had problems with accumulated spam, because of which supposedly letters from my domain accumulated in the queue and were not sent. This explanation was later partially confirmed by the fact that some of the belated letters still came. Since the mail was working and never failed again, my complaints about the mail were over.
')

Problem 2. Restarts

This problem manifested itself twice: according to my assumptions, the server with the database restarted for the first time since I received the message SERVER SHUTDOWN, for the second time my sites were not accessible via DNS, although access over IP remained, which is why I concluded that reboot the DNS server. The conclusions may be erroneous, but for the time being I adhere to just such an explanation of service failures. As for the time when the servers were unavailable, it amounted to a total of about half an hour. In the first case more, in the second - less, about 20 minutes and 5 minutes. If these figures explain something to someone, then I ask you to comment in the comments.

Solution of a problem. I had to wait, there was no other way out. Once again, these problems for the month and a half were only twice, they were not consistent, and the second case was short-term at all, so I did not go to the caliper for explanations and left everything as it is.

Problem 3. Plesk

While working in the admin panel of Plesk 9, which, admittedly, has a pleasant appearance and contains a lot of functionality, various problems have arisen. So, one of the main problems was the inability to create an alias for any of the domains. The process was simply interrupted with an error, although literally the day before that everything was fine.

Solution of a problem. Ticket was created in the caliper. The answer came quickly enough, a support employee said that he created an alias for me. After that, the creation process itself began to work normally and without failures. What could be the reason, and what the support staff did, remains a mystery, but, nevertheless, the problem was.

Other cases related to plesk, it seems to me, are on the conscience of the system itself, since errors occurred out of the blue: when navigating, when re-entering, when performing one sequence of actions several times, and so on. As far as I know, fixes are constantly coming out for plesk, so the new version 9 may soon become more stable.

Problem 4. Hanging

In fact, this is my problem, but I would still like to touch on it, because it shows how the support service treats such problems. The problem is that at the perfect moment all my sites suddenly died down and did not respond to requests. And on IP they pinged, ftp and mail worked. Obviously, the problem was in IIS, I could not draw any other conclusion. Something in the IIS server was hanging, and the query execution queue stopped, and then an error fell out. The hang timeout was very large, and the sites could hang for 10-20 minutes.

Solution of a problem. Two tickets were created, three times I talked with chat operators. The discussion in the caliper was very long, I insistently demanded the resolution of the problem, believing that the fault was on the hoster. However, the problem was that when the support service took up the job, and this happened after an hour or several after my request, they saw normally working sites, which I was informed about. Here the feature of this service manifested itself - they will not understand the problem more than is required. That is, the support service did not help me solve my own problems, reasonably believing that everything was in order on their part.

It is not known for certain whose fault it is and how much it lies only on me. However, after some changes in the database connection string, the problem disappeared. What made me think that I myself was to blame. As a matter of fact, maybe the hoster also did something and somehow reconfigured IIS, it’s impossible to know. Since the problem was real, I found it necessary to tell about it.

Conclusion

So, in general, that and all the problems that I encountered within a month and a half of using hosting. What conclusions can be drawn from the experience of solving problems:
- support is working properly, but not overly careful; solve their problems, but not yours; no one wants to help you if the hoster decides that the problem is with you;
- straight hands - the best medicine and before blaming the host, now, I try to find the causes of problems in my code;
- I estimate the total downtime of the web service due to the hoster’s fault at 25-30 minutes, the mail did not work once for about a day. The hoster promises a lifetime of 99.9%, which for 45 days means ~ 60 minutes of possible downtime. Without taking into account problems with mail, the hoster did it in the stated 99.9%.

As a result of working with the hoster, I want to say only one thing - I am satisfied. Even negative experiences, because he also teaches a lot. Mochahost satisfied all my requirements for technical capabilities, speed, support and even reliability, despite the comments. I continue to test further, do not finish, expect new articles.