Information Technology

As my current workstation only supports ‘slow’ PCI, there wouldn’t have been much sense in buying SATA-disks. Therefore, I have ordered a third hard disk of the given type. This will enable me to create 3x 78GB partitions which will result in a 156GB logical volume running on (SW-)RAID-5. On this I want to create a 120GB /home partition - this will be more than twice as much as I have today.

In addition, another 512MB of RAM and a USB-2.0/FireWire PCI card are coming soon.

For the server I bought two simple 20GB ATA disks, to be put on (SW-)RAID-1. This will carry the /home partition. As devil wished, one of the two SCSI-disks carrying the system crashed while building the two ATA disks in. So everything delays until the spare SCSI disk I ordered replaces the broken one.

Whenever sendmail has to deliver mails to other hosts which cannot be reached at that time, the messages are kept in the queue and are marked as “Deferred: Connection timed out”. Although the other hosts could be reached again and you want to tell sendmail to flush the mail queue, the command

sendmail -q -v

does not really try to reconnect to these hosts and still assumes that the connection timed out. The reason is that the hoststatus is cached, per default for a period of 30 minutes. Using

sendmail -OTimeout.hoststatus=0m -q -v

you can re-run the mail queue and force sendmail to reconnect to the hosts. You may want to define an alias for that, say, ‘sendmail-flush-timeouts’.

You can set further options in /etc/sendmail.cf.