PDA

View Full Version : Scheduled Maintenence: Mail server reboot 08/23/04 11:00PM EDT


allera
08-23-2004, 09:37 PM
This emergency notice was sent to all shared hosting customers:


Dear Customer,

The mail server, mail1.fluidhosting.com, has been experiencing abnormally high loads recently and we would like to run some tests and checks on it this evening (Monday, August 23, 2004 at 11:00PM EDT (GMT -4:00)). This will cause all mail to stop for 10-15 minutes. We understand this is a vital service to your account, so we are doing everything possible to keep the downtime as small as possible. We are also scheduling this as an emergency notification due to the situation, so it is not under our normal 24 hour advance notice policy.

Mail services should continue without problem until the maintenence window and we fully expect all services to resume as normal after the window.

Should you have any questions or concerns, please let us know by emailing support@fluidhosting.com.

Thank you,
Fluid Hosting, LLC
www.fluidhosting.com

Please note that during this downtime, all mail should queue at their originating mail servers for delivery. Once our mail system has returned to normal, all mail should be delivered normally. No mail should be lost. This is in accordance with SMTP policy and protocol.

If you have any questions, please email us at support@fluidhosting.com .

allera
08-23-2004, 11:27 PM
The checks and fixes have been run on the server and it is behaving normally again. No mail should have been lost in this process. Total downtime was approximately 10 minutes.

We also now have a second mail server in service, mail2.fluidhosting.com, which customers will have access to very shortly.

If you experience any problems with mail, please let us know right away by emailing support@fluidhosting.com.

Skibum
08-24-2004, 07:05 AM
What was the outcome of your discovery mission? Where was the high load originating?

allera
08-24-2004, 07:28 AM
There was a problem with the queue. The reboot was done to make sure the filesystem was sane before repairing the queue (fsck). There was also a particular non-existant email address that was being bombarded that was dealt with (I'm sure it affected the server a somewhat). However, the queue problem was the largest issue, by far.

All is well now. :)

Skibum
08-24-2004, 07:34 AM
Thanks! It's great knowing what goes on "Under the hood".