[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: So much for a wheezy install, massive fail



On Thursday 22 January 2015 09:42:13 Gary Dale did opine
And Gene did reply:
> On 21/01/15 08:36 PM, Gene Heskett wrote:
> > On Wednesday 21 January 2015 16:29:45 Bob Proulx did opine
> > 
> > And Gene did reply:
> >> Gene Heskett wrote:
> >>> Bob Proulx did opine
> >>> 
> >>>> You seem to believe that the Wheezy debian-installer does not
> >>>> handle the new Advanced Format 4k sectors.  However I use it all
> >>>> of the time with 4k sectors and it works fine.  There is no known
> >>>> problem using the Wheezy debian-installer to install onto AF 4k
> >>>> sector devices.
> >>>> 
> >>>> Sorry but I believe you are mistaken on this issue.  If you find
> >>>> the debian-install mishandling 4k sector devices please file a bug
> >>>> report with sufficient information to debug the problem.
> >>> 
> >>> Go ahead and install its way, then run an fdisk -l and read the
> >>> result, confirmed by quite slow readings from hdparm -tT on the
> >>> drive you just installed it to.
> >> 
> >> What problem are you seeing?  Details?
> >> 
> >> I might suggest that fdisk hasn't kept up and isn't the best tool
> >> for the task these days.  It may be getting confused by the newer
> >> partition tables.  This may be causing it to emit bad information.
> >> 
> >> In many ways I don't like parted but I think it handles the new
> >> formats best.
> >> 
> >>    parted /dev/sda unit s print
> >>> 
> >>> URL for the bug reporter?
> >> 
> >> Here you go.
> >> 
> >>    https://www.debian.org/Bugs/
> > 
> > Thanks.
> > 
> > gene@coyote:~/Downloads$ parted /dev/sdb unit s print
> > WARNING: You are not superuser.  Watch out for permissions.
> > Model: ATA ST1000VX000-1CU1 (scsi)
> > Disk /dev/sdb: 1953525168s
> > Sector size (logical/physical): 512B/4096B
> > Partition Table: gpt
> > 
> > Number  Start       End          Size         File system     Name 
> > Flags
> > 
> >   1      16384s      112656383s   112640000s   ext4                 
> >   boot 2      112656384s  215056383s   102400000s   linux-swap(v1) 3
> >        215062155s  317460464s   102398310s   ext4
> >   4      317460465s  1953520064s  1636059600s  ext4
> >> 
> >> Bob
> > 
> > Which it is not complaining about.  BUT that is not how I spent an
> > hour partitioning it last night, zero resemblance, partitions 2 & 3
> > were specced with  50G's for swap and /, the last, big one is /home.
> > 
> > Cheers, Gene Heskett
> 
> 50G for swap?!

The partitioner that sets this up previously setup a bit over 2x the 
memory, 18Gb for swap, then surveyed the system and found 2 more swaps, 
dutifully adding then to the /etc/fstab it wrote. So at that point I had 
nearly 50Gb of swap available in 3 pieces/drives.
 
> I also question having a separate /boot partition - especially one
> larger than 50G. I used to use one before mdadm RAID could boot from
> RAID 5, but these days I don't bother. It's just something that either
> wastes space or that can fill up and cause problems.

In a production environment, raid is ok if you can dedicate at least 5 
drives, we have an older centos setup at the tv station with its data 
array a 5 drive lashup, scary fast, but also quite complex to setup and 
maintain.  For some reason (they are very well cooled) the drive life 
seems to be limited to less than a year of 24/7 uptimes. So the last time 
Jim bought drives for it, he bought a 12 pack about 3 years ago. And they 
took that as a hint, and only 1 has failed since.  OTOH, the drive I use 
for amanda now has 

9 Power_On_Hours  0x0032 051 051 000 Old_age Always - 43534
on it, and has not recorded an error as I have its self-test enabled. So 
it would email me for a sneeze.  Except the test log ends at about half is 
spin up hours, so I wonder to myself when that was disabled. I just 
initiated a -t long since the drive will not see any activity again for 
about 14 hours.  I wonder it it ran out of buffer space at 21 reports?

Back on the thread..
 
Can linux now be booted as of old, with its vmlinuz file 3.5 Terrabytes in 
from the start of the disk?  That implies a bios a lot bigger & smarter 
than what was available at the time, circa 2007 or so, when I bought this 
Asus M2N-SLI Deluxe board, with a 2.1 GHz phenom, and built this machine.

Removing the boot partition removes the guarantee that boot related files 
will be within reach of the bios.  However, 50Gb is out of line as I have 
been using 1Gb for years, which has all sorts of cruft I haven't used in 
yonks in it.  So I'll likely fix that and slide the rest of it back out 
before I put another install disk in the optical drive.

Thanks Gary.

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>
US V Castleman, SCOTUS, Mar 2014 is grounds for Impeaching SCOTUS


Reply to: