12/05/03 mysql, mail, and cpanel servers

We had a power outage in the datacenter with one of the racks. I have brought them back again, they may be down for a brief interval after this as I am moving a few of the machines to another rack until we track down this power failure. The servers affected were:

mysql
mysql2
mysql3
mail
mail1
mail2
mail3
mail4
cpanel
cpanel2
cpanel3
ns2
ns3
redhatcomb

As I said, these are back up and running now.

[ark2000]

In the meantime there are some messages missing! Don't you have a UPS?

[Light Speed]

This is twice now in a week!

Please buy a better power supply for that rack

Yeah! Why didn't the UPS supply power while you switched out the bad PS?

The UPS's didn't fail, it was the cabinet that was plugged into it. The point of failure was in front of the UPS's. We are going to be working with our datacenter space provider to fix the faulty power hook ups in that cabinet.

I'm having to run fsck on cpanel.nocspeed.com and cpanel3.nocspeed.com. They should be up shortly.

And no mail is lost during this outage. If a message is attempted to be sent to our server and it is down at that time, the remote server will hold the message and attempt to deliver it until it can.

[ark2000]

Please explain what does the datacenter space provider have to do with your racks and your cables?

Do they provide them as well? Isn't that entirely your responsibility? The faulty power hook ups? If the UPS was plugged in the service provider's hook ups then the UPS should have kicked in, then it would have been your provider's fault but there would be no downtime.

If like you said the UPS is fine and the problem is the cabling on the cabinet, then unless you are leasing the cabinets from the same space provider I fail to see how can they be responsible?

Please clear this up.

I'm just trying to understand how this works... =)

[Light Speed]

Thanks for the update stalwart

[ark2000]

I'm tired of all those outages as well, I just downloaded:

http://www.hostalive.net/hostalive.html

and will set it up to monitor every site and server I'm using. I didn't want to bother with this before but there are too many downtimes lately, there may be some service refunds due.

cpanel.nocspeed.com is back up. cpanel3 is still running fsck. should be done soon.

I just moved mail.hsphere.cc. It is back up now. It was down for roughly 2 minutes.

[ark2000]

You are not going to answer my question?

It's simple ark2000, there is a big power strip in each cabinet. The UPS's are supporting the power strips and everything on it. The strip failed, therefore the UPS can't send power to the machine. They then shut off.

[ark2000]

interesting, after I posted the questions and comments above and installed the host monitoring application, someone at vortech decided to disable the ping on the mail server.... hmmmm

That's ludicrous ark. Read this thread, I am moving servers.

[ark2000]

you didn't answer my question about who's responsibility is the power strip?

answer my questions please.

It is not our strip, nor our cabinet, and as I stated before, we will be working with our provider to replace it.