On Thu, 28 Feb 2013 12:41:22 Russell Coker said
>I suggest that you firstly don't turn the system off until........
and a lot of good stuff...........
Thanks for the reply,
A comment on a couple of points raised, The drive that appears to be
failing only contains operating systems,ie no data at all, in this case
both WindowsXP and Linux. A long time ago I found this was an advantage in
managing data. I in fact keep spare drives handy (including an SSD) as I
live in a relatively remote area and travel can add quite a significant
cost to a system.
For most of my life I worked as a complex systems technician, and from the
__very__ earliest days was aware of the requirement to have good backs
_______AND_______ be able to restore such backup data reliably (I
highlighted the restore as a lot do not take this into account).
I in fact keep 4 copies of all data I consider to be of value as on two
occasions have had backups (other peoples) destroyed during restores.
The particular system is to be rebuilt next week so all should be well
then. The problem could just be a connector as I have had similiar problem
in the past with SATA power connectors not sitting properly.
Linux now for 20 years, (where does time go)
Lindsay
Many thank for the replies from both James Harper and Rohan McLeod.
The actual failure is rare and very likley is temperature related, failing
at low temperatures say below 20 to 25 degrees, it failed this morning for
the first time in quite a while, this being the first cold day in a LONG
time. The system shows no problems at all once it is running. All tests on
the hardisk, including thoses suggested by James being OK. The system will
compile a kernel OK and from past experience this is one of the best memory
tests on the planet.
From recent behaviour I think the most likely cause is the spindle motor on
the drive is slow in starting. I will though check __ALL__ connections as I
have had some __VERY__ strange things caused by both SATA data and power
connectors not being fully home.
BTW, the power supply tested OK, sudden changes in load effecting the
voltages not at all, allthough the CPU voltage was varing around 10 mV, not
worth worrying about though.
I am going to rebuild this system shortly (like next week, when my new case
FINALLY arrives) may just leave it switched on for now, its very unlikely a
thunderstorm will occur for awhile now. The power cost cetainly will not
break me.
Lindsay
PS, had around 75 mm of rain last night, man that was good to hear after
almost 5 months of nothing, nice cool day to so can get some work done.
Theres a fair number of people around NE Vic that will be glad to see the
back of this summer!!!!
I've got an old EeePC 900 that I want to put back into service. It's
currently running Xandros, but that distro is too old to be able to be
maintained or to be able to install most applications on. I'm looking
at replacing Xandros with something more modern.
What do people suggest as options? Most of the info on the net that
I've found seems rather dated (2009 or so)
--
73 de Tony VK3JED
http://vkradio.com
I need to move a block device (Xen VM backing store) between sites with a minimum of downtime... my choices so far are:
. dd - straightforward and foolproof but will take a while
. dd + lvmsync (https://github.com/mpalmer/lvmsync) - untested but sounds plausible
. drbd with external metadata - I already use drbd but haven't tried such a thing before
I think drbd is going to be the best way to go, especially as it's part of Linux these days. My plan is:
1. Stop the VM
2. create the drbd metadata in a separate lv and initialise it all (without overwriting initial data!)
3. Start the VM
4. create the backing lv and metadata lv at the other site and configure drbd
5. Pick a low-usage time and kick it off then wait for completion
6. Stop the VM at the origin site
7. tear down DRBD
8. Start the VM at the new site
Both sites are linked by openvswitch so they are effectively the same lan. Bandwidth is 20Mbits/second so it won't be lightning fast but the VM's are typically only 50GB in size which equates to around 7 hours + overheads.
Any comments? I'm about to test with a small and unimportant VM.
Thanks
James
Just wondering if anyone has any ideas on a problem thats just appeared.
One of my systems here ocasionally does not startup correctly. The system
is running Debian testing from around 3 or 4 monts ago. The Motherboard an
AUS P6X58D E Premium, drive layout is 1st drive an WD 160G around 6 years
old, 2nd an OCZ Vertex 4 128g SSD, the final drive an WD Caviar Black 1Tb.
The system dual boots with WIndow XP, windows C drive and the Linux root
drive both being on the 1st disk.
What happens in on startup the bios displays an id screen then displays
the mount of memory, then scans the USB ports for a drive, final scans the
SATA ports and then boots. When it fails it stops when it should be
displaying the amount of memory. Very occasionaly it will wait a bit and
boot with the boot loader lilo just displaying an L and the error number
99. This behavior could be instructive. The L displayed means the first
stage loader has successfully been loaded but failed to load the second
stage loader. The error number 99 is an unspecfied error. The system will
compile a kernel OK (2.7.4), although this only uses 1.8 gig of the 3.5
gig avaivible, it would seems the memory is OK.
It would seem the most likely suspect is the first hardisk, the 160g WD is
trying to fail. Although it currently shows no other symptoms, would this
stop the bios from proceeding? It could be instructive to read the 12 Volt
line going to the drive but I have yet to build an extension adapter
allowing my to cut into the power cable to the drive. I believe one can
purchase such an item now although I have never seen one as yet. Mind you
the PSU is a Corsair HX650 which is a known hi-quality item, although such
beasts can still fail. The drives do not fail under load as they usually
do when one has a dicey 12 Volt line.
Anyone have any ideas?
Lindsay
Hey,
Noticed something odd with my nginx setup recently, and can't debug what's
going on, so hoping someone has some ideas.
Deny just isn't denying. To test it, I've created a nice simple server
block:
server {
listen [::]:80;
server_name test.whatevz.net;
access_log /var/log/nginx/test.whatevz.net/access.log;
error_log /var/log/nginx/test.whatevz.net/error.log;
root /srv/test.whatevz.net/;
index index.php;
location / {
deny all;
}
}
And yet, as you can see for yourself, it serves up index.php happily.
I'm really quite stumped and would love some advice for what to do here.
I have another nginx setup where I've done the same test, and it denies as
expected, so there must be something local in my config causing this.
cheers,
/ Brett
I just upgraded a HP DL385 G2 from a 2.6.x kernel to the latest Debian 3.2.0-4 kernel, and now when it boots I just get this dumped to the serial port:
123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}
(eg the ascii character set)
The server is locked away in a data centre so I'm using serial console and can easily reboot into the previous kernel and it all works fine but I'd really like to get the upgrade done inside the maintenance window. Any suggestions? I've used the same Debian kernel on lots of other servers without a problem. I assume there is some kernel command line option I need to pass to get it going... I'm not sure if dumping the ascii character set is a symptom of a problem or a specific cry for help...
Thanks
James
Any pointers to Linux/X compatibility of the "Intel HD Graphics 4000" (as
used in Toshiba C850/0G6)?
Anyone had any experience with Linux on the Toshi C850/0G6?
Thanks,
Neale.
Running "X.Org X Server 1.12.4" with radeon driver 6.14.99 (being what's
included in Debian Whezy) on a new Toshiba L850/046 doesn't work,
apparently due to the driver not supporting the chipset.
lspci reports it as:
01:00.0 VGA compatible controller: Advanced Micro Devices [AMD] nee ATI
Thames XT/GL [Radeon HD 7600M Series] (prog-if 00 [VGA controller])
Any ideas on getting an X Server happening on this beastie?
Is framebuffer driver still a usable option in cases like this
(needing a workaround till a real solution is available)?
Thanks,
Neale.