20121020 another close call - plembo/onemoretech GitHub Wiki

title: Another close call link: https://onemoretech.wordpress.com/2012/10/20/another-close-call/ author: lembobro description: post_id: 3533 created: 2012/10/20 03:30:25 created_gmt: 2012/10/20 07:30:25 comment_status: closed post_name: another-close-call status: publish post_type: post

Another close call

The old Dell Inspiron 5150n came close to the garbage heap (or recycling center, as the case may be) yesterday night. For better or worse it's now back up and running, although I'm not sure whether to be happy about it. The "n" is for "no operating system", this 2005 era machine having been part of Dell's the short lived offering hardware with FreeDOS in lieu of the ubiquitous MS Windows. Sadly, the program ended as abruptly as it began. My 5150n has been running almost continuously since 2006, and the only components I've replaced have been the CPU1, DVD writer and the display adapter. Memory is maxed out at 4 Gb on a motherboard that doesn't really support addressing more than 3.5 Gb even with a 64-bit operating system. A 250 Gb Seagate2 was added early on to supplement the original 160 Gb Seagate boot drive. In short, this is an old machine whose final demise would be welcome on many levels. What happened yesterday was the usual confluence of failures that comes with old hardware. After several hours of debugging I came to the conclusion that both the replacement GeForce 6200TC video card (installed in 2006) and one of the SATA channels had given up the ghost. I retrieved the old ATI X600 display adapter that had shipped with the machine from a box in the attic and disabled the faulty SATA channel in the BIOS. Then I removed both the old hard disks and swapped in a single 500 Gb Seagate I had in my portable backup enclosure, connecting it to the still good SATA channel. From that point it was just a matter of inserting a Scientific Linux 6.3 Installation DVD and installing fresh on the "new" hard drive. Now comes the hard part: copying all my data from the old drives and configuring things to close to where they were before. Of course a lot of old cruft will get left behind in the process, that's the one up side to this kind of incident. Another of the many takeaways from this incident is the need to tighten up my backup procedures, and to archive state data such as the contents of /var/spool (in particular /var/spool/cron) so that it can be recovered. Having failed to include that kind of thing in my existing backups I'm dependent on getting it off the old drives -- not something that may be possible in a future situation. 1The CPU was a special case. The Pentium D 820 that shipped with the machine had been having temperature control issues and so I replaced it with a refurb Pentium D 945. With the CPU out of its socket it was clear that most of the temperature issues were due to the dessication of the thermal paste between the CPU and the heat sink. That could have been either the root cause of the overheating or a symptom of an overheated CPU. Needless to say I was extra careful to make sure I properly applied new paste on the replacement CPU (making sure there were no air gaps or bubbles on its thin surface) before buttoning the whole thing up again. The clock speed is only slightly faster with the new CPU and it still doesn't support the extensions required for KVM virtualization, and as noted above shortcomings in the motherboard design still prevent addressing more than 3.5 Gb of RAM even with a 64-bit operating system. 2I've come full circle over the years in the Western Digital vs Seagate debate. After years of favoring Seagate my disk hardware of choice is now WD, particularly the lower power "green" editions. As my middle brother used to say, heat is the number one enemy of all electronics -- and it is almost always true that higher power consumption results in more heat. Greener devices are therefore not just good for the planet, they should also last longer.

Copyright 2004-2019 Phil Lembo