Maintenance on the client mail server has been completed and it is restored to normal operation.

I will be taking the client mail server down for maintenance just after midnight to image it for recovery in the event of file system corruption.  This should take less than about 25 minutes.

During this time, pop, imap, outbound smtp, and webmail will be unavailable.  Incoming mail will still be processed.  Mail can be viewed and/or deleted using shell mailers like pine, mail, mush, mh, but outgoing mail can not be sent during this time.  Some mailers that use the local mail smtp server will queue, but any that try to send via mail.eskimo.com will fail.

The work on the switch upgrade has been complete.  Now even copying an 8gb file only increases ping times to about 65ms which isn’t noticeable where before it would lock them machines up for sometimes seconds because the file transfers sucked up all the bandwidth.

I will be upgrading the existing 100-base-T switch to a Gigabit switch to accommodate increased traffic between the file servers and other machines.

This increase traffic is largely the result of a move from NFSv3 to NFSv4.  NFSv3 wasn’t able to use the full bandwidth of a 100-base-T connection but NFSv4 will totally saturate the link during a large file move.  This is resulting in occasional latency spikes.

The change-out will create only brief interruptions of several seconds as machines are moved from one switch to the other (a physical movement of connectors over a distance of several inches).

Maintenance completed.  The web, ftp, and shell server scientific.eskimo.com are returned to service.

Just after midnight 12/20/2014 (tonight) [correction, since it will be after midnight, it will actually be 12/21/2014] I will be taking the web, ftp, and scientific.eskimo.com shell servers down for imaging so that we have current backups complete with all the recent security upgrades to restore from should it become necessary.

The expected downtime is about 25 minutes for web and ftp and about 45 minutes for scientific.eskimo.com.

Experience has taught me that nothing invites disaster like not being prepared for it.

I wasn’t able to complete all of the maintenance tonight although I did get the file servers and the busy shell servers and mail server.  Tomorrow evening after midnight I will need to bring the ftp and web servers down for maintenance for about 45 minutes each as well as scientific.eskimo.com.

I will be taking shellx.eskimo.com down.  It will bounce up and down several times as I make some partition changes and test to make sure everything is operational before backing up the new configuration.  The reason for doing this is to improve performance and simplify administration of this machine.  I expect the total down time to be around an hour.

The UUCP server uucp.eskimo.com and shell server debian.eskimo.com are going down for maintenance.  Expected downtime is approximately 20 minutes for UUCP and 45 minutes for debian.  After these come up the main file server will rebooted and everything will freeze for approximately 20 minutes.

Maintenance on mint.eskimo.com is completed.  It has been restored to service.