 |
Laboratory of Molecular Biophysics
Laboratory Journal 2004
R. K. Bryan
|
Return to:
Contents.
Computing
R. K. Bryan
Central Servers
The following systems provide central file serving and backup
services which are available to all systems on the laboratory network:
- A high-availability Linux fileserver cluster, comprising
- two Compaq Proliant ML370 G2 systems, running Suse Linux and
Mission Critical Linux Convolo.
- Jetstor III Raid array - six Seagate 180GB disks, 900GB
useable space, dual Ultra160 SCSI interface. This may be served by
either
system.
- Compaq MLS5026 SuperDLT robot. 100GB/tape. Data transfer
speed
up to 70Gb/hr. Magazine capacity 25 tapes.
- GForce RI Raid array with eight Maxtor 200GB disks, 1.2TB
useable space. Due to some continuing reliability problems, an
alternative stand-alone system is currently being used to serve this
storage, to avoid periodic reboots of the main systems.
- A VMS cluster consisting of an AlphaServer 1200 and AlphaStation
500, providing access to about 800GB of RAID disk storage, with a TL891
DLT tape robot. The storage is served to the rest of the laboratory
systems using various network protocols (VMScluster, NFS, SMB and
AppleTalk). This also runs network services for mail, POP, IMAP2,
IMAP4, DHCP and the laboratory web server. The inherent security of the
VMS architecture gives greater protection against potential external
disruption to these services. The operating system version was upgraded
to 7.3-1, which provides improved TCP/IP support, and the Samba
software (for file and print services to Windows systems) was updated.
- A 4-processor Compaq AlphaServer 4100 running Compaq Unix,
which serves about 50GB of storage, and acts as the NIS master node.
- In addition, two Beowulf PC clusters provide a high-performance
computing resource. The first, installed in 2001 for Mark Sansom's
group, consists of 32 dual 750 MHZ PIII nodes, and the second, a joint
resource for the Sansom, Noble and Garman groups which was installed in
2003, consists of 67 dual 2.4GHz PIV nodes, and also 1.8TB of local
RAID storage.
In general, all these systems have run reliably, with uptimes of many
months. Disappointingly, the economical RAID arrays using IDE disks
have not proved as reliable as their predecessors using SCSI disks,
although different brands and models of disks seem to vary greatly in
this respect.
Laboratory Network
The laboratory network continues to be based on a 100Mbit/sec Fast
Ethernet network, plus a Gigabit switch in the computer room which
provides a higher-speed connection to several of the central systems.
This network is connected via a Firewall (Intel PC running BSD Unix and
pf packet filter) to the University 10GB backbone ethernet, which
provides access both to other units within the University and to the
external Internet connection.
Part of Mark Sansom's group 'decanted' to the Biochemistry
building in the Autumn of 2003, and of course it was important that
they
continued to access the laboratory servers transparently, and in
particular the NIS and NFS services which are protected against
external access. This required a certain amount of reconfiguration of
our firewall, and for the Biochemistry Network Manager to make
corresponding changes to their firewall, plus a small amount of
retuning due to higher throughput.
About 200 host addresses are registered on the Laboratory network,
which include a wide variety of systems:- the central servers as
described above; a large number of desktop systems, including Intel
systems running either Windows or Linux, and Apple Macintosh computers;
personal laptop computers; and a number of systems dedicated to
specific tasks, such as control of the X-ray Area Detectors and EM
Image aquisition
Return to:
Contents.
Last updated: 26-APR-2005 14:26