[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Eskimo Server Upgrade - Outage July 31, 1999 1am-3am
- To: eskimo-announce@eskimo.com, outages-list@eskimo.com
- Subject: Eskimo Server Upgrade - Outage July 31, 1999 1am-3am
- From: Robert Dinse <nanook@eskimo.com>
- Date: Fri, 30 Jul 1999 01:26:14 -0700 (PDT)
- Resent-Date: Fri, 30 Jul 1999 01:26:19 -0700
- Resent-From: eskimo-announce@eskimo.com
- Resent-Message-ID: <"OUQEN3.0.zO7.g6Met"@mx1>
- Resent-Sender: eskimo-announce-request@eskimo.com
Eskimo North will be off-line I am guessing for about two hours
starting at 1am on saturday morning for a hardware upgrade. Because the
eskimo server is so central to everything else, this will affect many
things. Incoming mail will not be processed during that time frame, it
will not be possible to send mail during that time. The shell server will
be down, and you will not be able to login to the low speed modem pool.
What follows is mostly a technical explanation of those upgrades and
the progress of other upgrades under way.
We are replacing the 4/670MP with an SS-10. This provides a disk
controller twice as fast as the current machine and a faster bus with a
faster memory subsystem that will improve overall system performance. The
video card, fddi card, and disks are being re-used from the existing
machine so there is a bit of hardware to transplant. There is the remote
possibility that the upgrade will go smoothly and the outage will be
shorter but rarely do these things go entirely smoothly.
On other subjects, Eskinews hardware is upgraded, we are still
working on upgrading software. The kernel is 2.2.11, though I am going to
go back to 2.2.10 because performance with 2.2.11 is somewhat worse.
The machine is mostly idle, both disk and CPU, feed from Sprint is
abysmal, and I am working on a fix for that involving a feed from another
company. Assuming we can get the feed fixed, and the machine provides
adequate thruput, then I will add additional disk for a larger spool and
longer expire times.
I am still waiting for the machines that will replace chat and
isumataq, these should arrive friday or monday.
I also acquired another router with faster CEC (the central processor
unit), and more ports, and most importantly a 100-base-TX card. FDDI has
not been supported by SparcLinux, and that has been holding us back from
upgrading some of the machines to SparcLinux, some like mail could really
benefit. While not as fast as FDDI, it will probably provide adequate
bandwidth. I am now looking for a reasonably priced 100-base-TX switch.
Finding a way to get from FDDI to 100-base-TX has been difficult, this
will provide that.
This will provide more CPU and more ports as well, so that we can add
additional bandwidth (which actually is something I have in the works).