This was originally going to be a very long post, but I kept putting this off ... and now I just feel that something needs to be said.
The story starts with waking up on February 15th, to find zen was dead. It had self updated overnight, and now it was unbootable, and the start repair couldn't get me back. Apparently, the problem had started long ago with all the previous times where Windows 7 would lock up...usually under intense disk activity...and the afterwards, the intel matrix raid would require re-initialization of my 1.5TB RAID 1 array.
Apparently, it was slowly corrupting my drive....because trying to restore from WindowsImageBackup was also a failure. Since this happened the day before Gallifrey One, I had to wait until I got back to do some more serious attempts to recovery, during which I ordered a full copy of Windows 7 Professional, hoping that a repair install might be an option. It isn't because the repair option can only be invoked inside a running Windows 7 system .... in need of repair. Not by booting the disk. ARGH!
At least I should have the data in BackupPC to restore from.....though hopefully before the bit rot of its ext4 filesystem makes it go away. Plus I had hoped to get some configuration going where I could mount the RR62x RAID 5 array, and get at the Oops!Backup store.
So, the plan now was to wait until Ubuntu 12.04LTS to land and then maybe some configuration of running Windows 7 in VirtualBox and recovering into that, etc.
Well, I was getting really annoyed with it complaining and kicking out the replaced disk in LHAVEN. So, I shutdown the system and took out the disk, figuring it should just continue to run fine degraded until I get around to replacing it.
Well, it wouldn't power on after that. I went through everything, no joy. It had done this a while back when I tried adding something to it. It could just be the power supply has flaked out, but I don't have a tester or multimeter handy...so I can't really test it. Could buy a new PS...and I may do that.
But, I kind of suspected the drive issues weren't actually the drive but possibly some deeper hardware problem. So, I had been planning to replace LHAVEN at some point. It has done quite well, being circa 2002. Started out as a 64MB Duron 800MHz machine, eventually peaked at 2GB Athlon XP 3000+. It had replaced a Cyrix PR233 box that had gone up in smoke during a hot summer day in 2002, when the transformer outside blew....the computer survived the brownout and then blackout, but the fans didn't spin back up when power returned...so it burned itself up. Later I found one problem with this new system. It wouldn't resume after losing power. No BIOS setting to alter this behavior and I did try to see if there was alternate BIOS updates for it. It was kind of a painful machine to manage, because for some time kernels didn't have built-in support for all the SIS chipset stuff in it....so an update would roll out, and NIC was usually one of the things to definitely go missing. Which made for fun to rebuild a custom kernel upgrade for it. Eventually it got stable.
But, needing essentials like DNS, DHCP to be available after an extended outage...I moved these to another server (originally an old Pentium 75, which has evolved into what is known as 'box' today). For the longest time it was RedHat 7.2, and then RedHat 7.3 when fedora legacy switched to only supporting 7.3 and 9 releases. It continued after fedoralegacy stopped supporting it, I was building some of the packages for it by hand...like bind (in response to the Kaminsky exploit). There were parts of my network that wouldn't function without LHAVEN...like an old Windows 2000 box that was barely functioning...(old Gumby).
I nearly lost it during the Icepocalyse....but I tracked down motherboard replacement for it from geeks.com. It was a slighty newer mobo, but equivalent chipset. It did have some things the old mobo didn't, like USB 2.0 support (I didn't use USB and still don't), support for 2GB of RAM instead of 1GB...so I upped it for better BOINC'ng. And, faster CPUs....so I upped from 2200 to 3000.
At one time, I had turned it off to do some upgrades...which didn't pan out. Tried to slap a gigabit card into it...didn't work. Wouldn't get along with the on board stuff, and the BIOS didn't have ways to get things out of the way enough. It also never fixed the resume after power loss issue. But, during this process it failed to power on....but after a couple days, it came back and I decided to upgrade its UPS and hopefully it would make one last outage. Well, there were several others since then where it came back afterwards. But, looks like this is the last time for it.
Pages: 1· 2
...was less than successful.
Back following the icepocalypse, I resurrected 'lhaven' by changing its motherboard. The new motherboard was an upgrade to the old one, but I wasn't taking advantage of any of it. But, in my mind I thought that someday I might.
Having finally upgraded the OS from RedHat 7.3 to Ubuntu 10.04....and being able to participate in all BOINC projects again, I decided that I might look at upgrading it finally.
So I went looking for the fastest FSB 333MHz CPU that I could put into it. And, went looking on eBay for one. After a few unsuccessful attempts, I went and bought an Athlon 3000+ (and an Athlon 2400+). And, while I was on there, I got a new CPU Heatsink and Fan and a 1GB stick of DDR333 memory.
I wasn't sure what kind of memory was in the machine, but somehow I thought it to be a 1GB stick of DDR333 (since I for some reason had an extra DDR333 1GB stick that I gave away recently)....but memory told me the machine had two slots, so I'm not sure why I didn't upgrade the memory to more than 1GB...
Well, turns out my memory was faulty (as was lshw)...the machine had two 512MB DDR226 memory sticks. Which I makes sense, since the old CPU was FSB266 (a 2000+)....guess that means I shouldn't have given away that 1GB stick of DDR333.
What I did was upgrade the memory to 1.5GB by putting the new 1GB DDR333 stick in the first memory slot, and a 512MB DDR400 stick in the second slot. I then looked at the CPU.
First I put the Athlon 3000+ in....and the new CPU heatsink/fan. It is a touch bigger, actually its a lot bigger than the old....not quite compatible with my case...though it should be a simple enough mod once I get everything working again. Except I didn't. No POST.
This year I'm taking an more active roll in the organizing of it, then I did last year. And, it is shaping up to be pretty interesting. Though budget wise, its not looking like I'll be able to make....but now I'm thinking I might scale back my travel plans so that I can still make it.
I'm, again, primarily helping review speaker submissions for who will get to present. And, some of the people already lined up as keynotes and such seem quite interesting.
For instance, Peter H. Salus will be a keynote at OLF this year.
Also Shawn Powers had been invited to be a keynote. Wonder if this means that I'll finally get to meet him in person. He was at Penguicon earlier this year, but I only saw him twice...once as I was rushing to catch one of the elusive slow moving elevators. And, the other time while I was in the beef line.
Among the other chatter about possible speakers...is Dan Kaminsky, which stands out as a name I recognize because of the DNS exploit and all the extra work that got dropped on my lap last summer. Though later it was fun building new bind for my home DNS servers (the RedHat 7.3 and SuSE 9.3 boxes).
Elsewhere...Maco commented that if I pick her submission, there'll be at least one system administration talk at OLF. And, my response was that I don't know if I want to hear system administration talk when I'm on vacation. Though the whole idea of OLF for vacation is kind of odd. Especially since stress management says I need to go on vacations where I completely stop thinking about work.... Guess we'll see when I start reviewing submissions.
The last time I was ever on such a vacation...was in 2000. There was enough activity to keep me occupied and there was no Internet access or cellular service out there. Both of which are now available (which impacted the trip I took in 2005...in fact the page of services down just as I was departing really messed things up...and the it wasn't fun trying to do remote access on the spotty wifi/satellite Internet service.
There are also other features at OLF that I'm interesting in seeing/hearing. Like the DIOS panel on Sunday. (Diversity In Open Source). Or hearing Elizabeth Garbee speak again, which turned out to be rather interesting to hear when I heard her speak last year.
The linux server known as 'box' died from the rain storm on Monday....well, the powersupply had been iffy for a while, and guess it toasted when things came back.
Sounds familiar....since its predecessor died in a similar manner....transformer blew....but that was fine, I was protected. And, then its UPS ran down...and all was quiet. It was also like 100 degrees that day....so it was pretty miserable. But, when power came back...it resumed like a good server, but the fans in it didn't. So it burned up.
Something like that probably happened to 'box'. Actually the predecessor was before 'lhaven'. The problem was I later learned that the replacement's BIOS had no disposition on what to do when power is cycled (no stay off, turn on or last state option). It always stayed off. The old 'lhaven' was a PC-AT system, while the new 'lhaven' is ATX. (post icepocalyse this problem was fixed).
So, after losing core services due to a power outage leaving 'lhaven' off....I decided I needed a new server to run these services on (DNS, DHCP, NTP). I didn't have to be much, so my then employer was having an auction of old computers....and I picked up a Pentium 75.
Wow, that's slower than the first PC I ever owned (before PC I was a Commodore person). I had acquired it with the plan to run ShareTheNet or some other Linux router thing...this was before Sveasoft and the other options became available for the WRT54GS....which is what I do now.
But, I installed SuSE on this P75 and was in business. Over the years, the hardware has changed. Actually it might have initially been RedHat, I did have a harddrive fail (they were 40G drives from ReplayTV upgrades). But, the current incarnation is SuSE 9.3 based.
Along the way there was a newer off-lease Dell that I acquired from eBay....move drive, let it rediscover hardware and make it go. Then that computer failed. So, I got an off-lease IBM NetVista from TigerDirect. Drive transplant, make it go.
By this time I was getting fancy with the stuff I was doing on it. At first it was be an internal router for my ReplayTVs, so I can sniff the progress of IVS transfers. Later it was run pulsed so that my Sprint 8703e could be bluetooth tethered. (something that I had taken for granted with my T-Mobile 7200....used tethering a lot during the first month I moved to Manhattan, KS...because I was having trouble getting Cox....missing the a few feet of coax between the pole and the box outside).
There are/were other things on here...aside the obvious of BOINC, there was ddclient and it was also the Linux box with a CD burner (for when I don't want to walk to my Windows computer and use the CD or DVD burner in there). 'orac' has a DVD burner, but I haven't used it yet.
Anyways...the question is will I get 'box' working again (either by fixing its hardware, or move the harddisk to a new machine and make it go again)....or make a new 'box' and recreate the core services. I no longer need pulsed (even though I don't have tethering ability with my iPhone, and I don't plan to pay for it....tethering is something I only do in a pinch, so I don't want to pay for something I hope to never need. At least with voice minutes those rollover, even though I hardly use them still.)
There's a certain temptation to load ubuntu and run these services, and then be able to keep the box current (more easily, it was fun building bind a couple of times....) Or hacking the system to support the new Daylight Saving Time.
OTOH, I kind of liked that the machine rather changed for something that was supposed to just be like an appliance.
Hmmm.... There is Ubuntu 8.04LTS, this is what 'orac' is running....ya, that's what I might do. Doesn't matter about the other versions of Ubuntu that I also run here and there. Having them all the same isn't an issue.
I guess it largely depends on whether I go with a different computer or fix the old computer. Both options are on the table now...since I have another IBM off lease computer on order (a ThinkCenter S51, IIRC and a powersupply [pull, untested] from eBay.) I think the 'new' computer will arrive first though. Though it'll probably be sometime before I get a new system up to the same level that 'box' was.
Oh yeah, 'box' was also an SMTP server....had moved that from 'lhaven' during its downtime following the icepocalyse. That reminds me....need to see which mail server 'tardis' and 'lunatic' are pointed at....'orac' was pointed at 'lhaven'....had found this morning that 'gumby' was stuck trying to send mail. (wallwatcher output to DShield)
I'm starting to think now that I should've held out for the ASUS Eee PC 1000. I'm finding the 901 a bit too cramped in physical size and storage.
I had thought of blowing off the Xandros and doing Ubuntu, which would probably help with space by not having all the stuff that is bundled. But, I opted to stay with the Easy Mode/Xandros. Largely because I'm looking for the device to just work.
I did restore it a couple of times to factory settings, because I took a blind turn playing around with apt-get and ended filling up the root drive and making a general mess of things.
Though I do have an issue with its wireless....it seems to take forever to reconnect to wireless after coming out of standby, and, then more often than not, it then can't connect anywhere. Even though all the settings are right. Have to stop and restart the wireless to make it go.
It'll be a few days...maybe a couple of weeks before I get it all customized to fit the way I want to use it. Partly because I know what I want to use it for, I just don't know completely how to get there.
Compounded by the fact that it is replacing my previous laptop which is dead, and I left my work laptop at work this weekend (and I don't know what next weekend holds). But, I'm not desperate enough to keep moving forward on getting the Eee PC customized to go into work to fetch my work laptop.
But, I have did already crack open the unit. I upgraded its memory to 2GB. Only hitch was getting the new DIMM to fully seat. I have a slot loading USB External DVD burner on order for it...could be a while, its coming from a Hong Kong based eBay seller. That reminds me...I have something else coming from Hong Kong that I'm waiting for.
I'll probably even be on to my other gadget projects before I come back to fiddling with the Eee PC some more. Basically, I had wanted it before classes started, but now that classes have...it may be a while before I can get to it. And, I'm sure by then...the 1000 would've been attainable. The extra 4GB (to 8GB) of the primary drive would've surely made a huge difference on how customized I could go with my stuff. And, I'm sure the extra space of the secondary drive (16GB to 32GB) would've gotten used for something.
Even though the plan is to avoid storing anything long term on the device. Though I'll probably still want some kind of backup plan for it...because it'll be a pain to lose things once its customized.
This has been a hot topic lately....which I first became aware of when an urgent ticket was assigned to me on Tuesday to upgrade all the campus DNS servers to a 'safe' version of bind.
During the winter break I had updated the campus caching DNS servers to 9.4.2, but the primary/secondary DNS servers were left running 9.3.4.
The 'safe' versions were 9.3.5-P1, 9.4.2-P1 or 9.5.0-P1.
The main campus caching servers had been running 9.2.3...so they were more pressing to upgrade to 9.4.2. During the break, the data center caching DNS server died...which created a good time to build the latest 9.4.2 version and later I upgraded (and make consistent) all the caching servers to this version.
In response to the urgent ticket, I got all the caching servers to 9.4.2-P1. And, this afternoon I got the primary and secondary DNS servers upgraded to match. I also discovered that the secondary DNS server had inadvertently been acting as a caching server to the whole world. After I turned it off, there were lots of log messages of comcast addresses being denied cache queries....and it was for a variety of well known sites, including youtube, paypal, google. At first it was Michigan comcast, but later I saw numerous other states, such as FL, GA, CO....
Meanwhile...I had started looking at getting the new bind package onto my Linux servers. These servers are well past EOL. So, I knew I was on my own to get things work.
Anyways....I was able to get bind-9.4.2-P1 to build and eventually run on both my RedHat 7.3 server and my SuSE 9.3 server.
Guess I need to update my ubuntu (8.04 server) from its desktop....
Latest Poopli Updaters -- http://lkc.me/poop
|<< <||> >>|
«windows xp» «watch instantly» «powersource 400» lhaven 10.04lts appletv tardis ups «hd movie» boinc upgrade amazon.com mdadm «air purifier» eyeglasses «tivo premiere» dvd replaytv orac cfengine3 usb ubuntu dsl voip ebay zen tivo «sans digital» cox progressive b2evolution twitter «doctor who» cpap freebsd «instant streaming» «tivo hd» backuppc prescription «windows 7» «chicago tardis» woot raid «amazon prime» netflix tv raid1 virtualbox linux box