Hello,
Apolgies for the delay in this summary, but organising the training takes its
time :-) The original questions were:
> I am about to start a little bit of Sun administration training for our Ops
> staff. I am not worried about the main part of it, but am sure that I'll be
> asked those 'awkward' little questions to which I will be stumped :-) So,
> in anticipation, I've been trying to think about what *I* don't know (okay,
> lots I agree...), anyway I ended up with 2 little questions:
>
> 1) When the system (solaris 2.5.1/2.6) is shut down it syncs the disks and
> says
> syncing disks...[2] [2] [2]
> or some such thing. The numbers may be different, but what are they?
> What do they represent?
>
> 2) When formatting a disk the 3rd slice (eg sd2c) is left as 'the whole
> disk'. I have only been able to find out that 'this is historical'. But
> what is it used for (or was it used for)? And any reason why it has to
> be the third slice? (I'm sure I looked this up before somewhere, but
> can't remember where. It's nothing to do with the mapping of bad blocks
> is it?)
>
I received a few replies and quite a few requests for the summary!
Thanks go to:
"Arthur Dunham" <add@netcom.com>
"Jim Harmon" <jharmon@telecnnct.com>
"David Schiffrin" <daves@adnc.com>
"Sean Ward" <sdward@uswest.com>
"David Thorburn-Gundlach" <david@bae.uga.edu>
"Chris Drake" <Chris.Drake@corp.sun.com>
"David Evans" <djevans@au.oracle.com>
"Davidson, T.(CIV)" <davidsot@scolan.centaf.af.mil>
The answers:
1) Chris Drake wrote:
> There are two sets of numbers: those in brackets and those which
> aren't...One refers to dirty pages, the other to dirty inodes (I think),
> and I can't remember which is which. Essentially it's a counter for data
> being flushed to disk. If it says "done" that's a good sign. If it
> says "giving up" that is not super, but not catastrophic: means that the
> system was not able to flush everything; you may have some lost data or
> filesystem damage to repair.
David Schiffrin wrote:
> The numbers are the number of blocks/buffers remaining to be flushed
> to the current cylender group . As I recall, there's a bug in the
> calculation of these numbers for display in (I think) 2.4 through
> 2.5.1 I don't know if it still exists in 2.6, but the nature of
> the bug is in some instances the number never decrements.
Thanks to others for replying to this as well.
2) The 'c' disk slice is indeed historical. It use to represent the entire
disk when backups were more 'limited'. This enabled the entire disk to be
backed up in one go - a 'raw backup'. Nowadays this is not generally
necessary, and the partition can actually be used if necessary.
Thanks to you all.
John.
-----------------------------------------------------------------------------
John Horne E-mail: J.Horne@plymouth.ac.uk
Computing Service Phone : +44 (0) 1752 - 233911
University of Plymouth, UK Fax : +44 (0) 1752 - 233919
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:12:40 CDT