Account suspended. Suggestions needed...

[The Grinch]

What about phpMyChat? It's available to install right from cpannel.

[Pedja]

We had quite loud discussion just before old forum pased away about this issue. Unfortunately, Andrew does not have understanding for such problems.

Quite often people complain that their cpanel acounts were suspended for good because issues beyond their control.

It is already concluded that on Dathorn we are all sitting ducks. If someone wants to hurt ou all he hs to do is create large traffic with your site, which is easy to create. DAthorn will suspend your acount instantly and you are out of the business.

A week ago I was close to the edge to loose tolerance about issues Dathorn caused to me, but I managed to control myself. I am not a weeny, and i said no word about problems Dathorn had in last fe weeks. I thought if I cannot help I should not make tension.

But after that, cpanel went crazy and most important account was partialy damaged. I had full backup saved just before crash so I was not worried. I asked support to help and they offered to recreate account from my backup. After their intervention, account ws totaly devastated. I got explanation that my backup was not correct. That was backup created by cpanel, and that was best backup I can create. Support comment was that cpanel backups are known as incorrect. Of course they said it after restoration failed. If they knew it may fail they should ate least say so before.

However, they offered to restore account from their two days old backup. I agreed but after that I had same mess. Their backup was not of any help. After I asked what is then backup procedure we have to do to filfil Andrews requrements to create backups (whenever someone mentions that he lost some data, he answers that backup should be made) I was left without clear answer. Andew said that I should ask them to create backup for me, but did not explain why their backup would be better than one they already had and which was unusable. I was quite pissed off. Clients were jumping over my head and I could do nothing and even support could do nothing, so I had to turn of my phones, crete account from scratch and wait system to propagate settings before turning phones on. I probably lost few clients because of this, at least they announced so.

I do not say Dathorn supports is bad. On the contrary, most of the time they do great job instantly and I always recommend them as hosting service and I do not think of leaving them. But I do expect them to be more tolerant on issues that are beyond control of account holder. They should think of what would happen if we became as untolerant as they are, on failures they control and even those that are beyond their control.

I am seriously concerned that this what happened to you and other may happen to me, and that I would not be able to fix things except leave Dathorn. My concern is real. I am involved in antispam activites and, often, I am spammers target. If they find out this, I will be out of business in no time.

[andriven]

Depending on the version of "ps", either

ps -xa | grep mysql 

or

ps -xam | grep mysql 

should list all the mysql processes running...one line per mysql process. You can then pipe that output into "wc" for a nice easy line count.

ps -xa | grep mysql | wc -l 

or

ps -xam | grep mysql | wc -l 

If you want to really keep track of it, you could cron it to run once an hour with a date stamp and append to a text file. Something like these lines in your cron tab.

00 * * * * /bin/date >> /home/username/mysqlcheck.txt
01 * * * * /bin/ps -xam | /bin/grep mysql | /usr/bin/wc -l >> mysqlcheck.txt 

Only the offending account was locked down. But in reality that is the only one I care about. I'm not reselling accounts. I run a local car club and dathorn looked like a good home for hosting it.

I was not given warning about the issue, but from what I have seen in the past, things are run on a zero tolerance policy here. So I'm not exactly shocked at the outcome, just shocked that it happened. The site was using about 22GB of transfer per month. So as you can see, the board was fairly active. The issue here seems to be a CPU load issue, not a transfer bandwidth problem. I don't know of any way that I can monitor that myself.

What is interesting is that I've moved the forum (the most used part of my site) to my home test server. It's a puny little PII-266 with 192MB of ram, running linux, and a home cable connection with an upload speed somewhere between 30-40KB/sec. So far everyone has told me that the forum seems to run fine, and isn't all that slow considering the bandwidth restrictions. So if it's running on such a small system without a hitch, why it was crippling the dathorn server is a mystery to me. My only thought is that it was a glitch, some kind of web attack, or I was being spydered big time. I really wish I new what it was so that it doesn't happen again anywhere else.

How would I be able to see how many MySQL or PHP processes that are running? I'd like to check this on my own server for reference.

[andriven]

You can also use "top" and "uptime" on your linux box.

uptime tells your uptime and how the load on your box (load = 1 means that there is always 1 process waiting to run). You can "man uptime" for more info.

top gives you a breakdown of all processes and which are chewing the most CPU time. "man top" as well.

[andriven]

After I asked what is then backup procedure we have to do to filfil Andrews requrements to create backups (whenever someone mentions that he lost some data, he answers that backup should be made) I was left without clear answer.

For what it's worth, I have a set of scripts that I use to

1) dump all mysql databases to text files nightly
2) keep the last 3 mysql dumps (script rotation)
3) use rsync (very bandwidth-friendly) to backup my home directory once every six hours to a home Unix box (actually Mac OS X...but the scripts work on any Linux/Unix as they're written for the sh shell which everyone has)
4) use the power of hard links in cp to keep rotating backups on the Linux (3 daily, 3 weekly, 3 monthly) in only about 2x the space of the home dir.

I'm working on cleaning them up so I can sell them for $15-$20 on WHT (people buy the scripts, modify some variables and run them). If you're interested, let me know.....I've been using them with good success for a while now....so nice to have really good backups in not much space.

[DomainDog]

But I do expect them to be more tolerant on issues that are beyond control of account holder. They should think of what would happen if we became as untolerant as they are, on failures they control and even those that are beyond their control.

I think the most important point Pedja has made here is- Andrew asked all of us to be very patient while he and all of Dathorn experienced a huge crisis that was out of his control, yet affected every one of us. To my knowledge, while Andrew rectified the problem and made everything good again, most everyone WAS patient and understanding, and endured the upset and the disruption with great loyalty and understanding- to their eternal credit.

Should this not be reciprocated? Especially in cases where something is beyond our control?

Where would Dathorn be if all the disrupted customers had filed a class action lawsuit against Dathorn? Where would Dathorn be if The Planet had decided he had violated some clause in their TOS, and shut down 100% of the Dathorn servers? Would he feel this was justified, or would he be voicing the same concerns that have been voiced here, but on a larger scale? Would he have requested a second chance?

In most cultures and in most businesses, tolerance within reason is considered a virtue. A wise philosopher once said, "From those to whom much is give, much is expected." Perhaps there is a lesson there for all of us.

PS- another example of an account suspended with no apparent explanation: http://forums.dathorn.com/showthread.php?t=106

[The Grinch]

I type in

ps -xa | grep mysql | wc -l

and it returned the number 6. This was on my home server which is now my temporary host for my forum. That seems quite a bit different from the 800 processes that were apparently running when my account got suspended.

I dunno. Something just doesn't jive.

[AndrewT]

There is no point in arguing this, our stance on this is not going to change. Sure, we can be patient and allow you to sort things out that are out of your control, but how do the other 50+ cutomers on the server feel about this? There is a big picture here that is much more important than any one domain/customer that may be causing problems and that's the bottom line.

[The Grinch]

I'm not asking you to change your mind on anything Andrew.

Anyway, I re-ran

ps -xa | grep mysql | wc -l

today and now it says 7. What would cause that to increase from 6 to 7 over night? Is that normal?

[andriven]

If everything is running correctly, the number of mysql processes just depends on how many people are hitting the board at the time you run that command.

If a mysql process gets stuck, that number will likely climb and climb....but otherwise it will fluctuate based on the number of active users (one active user might generate more than one mysql process depending on how the board is written).

I'm not asking you to change your mind on anything Andrew.

Anyway, I re-ran today and now it says 7. What would cause that to increase from 6 to 7 over night? Is that normal?

[KyleC]

i am going to split this thread its becoming a mess

here is the other half: http://forums.dathorn.com/showthread.php?t=150