You have decided to buy a SPARC based workstation, now you can tell what CPU each model sports, but can you tell how that particuliar model looks like? How its is inside?
Do not worry, this section will help you on that topic as well as providing you with some hints on how to inspect and test the hardware.
PC hardware is everywhere and usualy one is quite familiar with it, this is not exactly the case with SPARC based hardware, even more when it comes to the innards of a workstation. The good news is that it exists some sites on the web, where you can find pictures of Sun hardware, with some very detailed shots, thus you should be able to instantly identify the model and its condition prior to buy it. The two main site where I usualy go are:
On the one hand, it is very easy to find information on Sun hardware, while it can be difficult to find it for clone systems on the other hand it should be more fun to work with exotic hardware.
Clone systems have been or are still manufactured, at least by:
Of course if you buy first hand hardware everything should be fine, but if you decide to go for second hand hardware, you will have to decide how you want to buy it, and you should be able to test it by yourself.
Second hand hardware can be brought from Sun as refurbished hardware, at auction websites, or in specialised stores, or directly from companys that upgrade their hardware.
When it comes to second hand hardware one has to be as cautious as possible about the source, and the condition of that hardware.
For the source of the hardware, if you can track it, it is a very good point. If you can't track it, it is up to you to decide if you trust the seller or not ( If the seller cannot give you the reference of the hardwre and if, obviously he/she is clueless about Sun hardware, you should switch to condition red ).
Another interesting point is to see if check is accepted as a payement.
If you can check the hardware before buying it
then first have a general look at it, search for
cracks, for stains; check the connector's pin. If this first
inspection is OK, then ask to see the inside, look for spills, watch
carefuly the connectors, then if it is OK ask to see it running,
watch carefully boot messages, issue the dmesg|more
or if it runs Solaris you can issue the more /var/adm/messages
command,
and also very important listen to your computer, do you hear unusual
noises? Does it smell OK?. Then enter the OpenBoot by the
stop+a
command and run some tests ( see the OpenBoot section
).
When you have decided to purchase it, it is very important that you
always make sure that the workstation you are purchasing is the one
you have tested: always keep an eye on it, do not let someone go to
the backoffice with it for example, do not accept to leave without
your workstation.
Then, when you are at home, recheck it as if it was the first time you see it. If it comes with CDROM drive, try to mount/umount a few CDs and read them in order to check that device. If there is a floppy disk read/write/format a few floppys, this should be a good test.
Of course, it exists tools to automaticaly test the hardware, but usualy you do not have them when you need them, thus the script below relies only on ressources provided by a Linux base system.
You should run the following script for three days, basicaly it is going to use the CDROM, floppy and hard drive, and to reboot the computer every 3 hours. This should stress it,
# # Enter this in the crontab(5) # run the torture.sh script every 3 hours if possible # 0 */3 * * * /root/torture.sh
torture.sh
is
#!/bin/sh # # these are for controlling the loops # the CDLIMITS and FDLIMITS # are for the CD and FD loops # CPT=1 CDLIMIT=3 FDLIMIT=10 # # to which devices are # CDROM and floppy attached # CDROM=/dev/sr0 FLOPPY=/dev/fd0 # # where is the program we # intend to compile PATH=/path_to_big_package_to_compile # # this section si meant to # test the CDROM and floppy drives # comment what you do not need to # test while [ $CPT -le $CDLIMIT ] do # # CDROM drive # mount -rt iso9660 $CDROM /cdrom find /cdrom -exec cat {} \; >>/dev/null 2>/dev/null & umount /cdrom CPT=$(($CPT+1)) done # # floppy drive # CPT=1 while [ $CPT -le $FDLIMIT ] do mke2fs -c $FLOPPY >/dev/null 2>&1 mount -t ext2 $FLOPPY /floppy cd /bin/ cp dd ps echo sh /floppy find /floppy -exec cat {} \; >>/dev/null 2>/dev/null & umount /floppy CPT=$(($CPT+1)) done # # The big, intensive # compilation # cd $PATH # # now compile # make # # # we remove every .a .o .s and every executable # find $PATH -name "*.[aos]" -exec rm -f {} \; >/dev/null 2>&1 find $PATH -exec test -x {} \; -exec rm -f {} \; >/dev/null 2>&1 # # time to reboot # reboot
Of course you should run this script once in order to know how much time it takes to complete, this is just an idea on how to automate things.
Закладки на сайте Проследить за страницей |
Created 1996-2024 by Maxim Chirkov Добавить, Поддержать, Вебмастеру |