On Thursday, September 22, 2011 01:43:56 PM Charles Curley did opine:
> On Thu, 22 Sep 2011 09:39:12 -0400
>
> gene heskett <[email protected]> wrote:
> > Greetings all;
> >
> > I disturbed amanda's schedule about 3 weeks back by shutting the
> > whole system down for 6 days while the wife and I took a few days to
> > "get out of Dodge".
>
> The only reason I can think of to shut Amanda down is because you are
> changing tapes manually. I gather that you are not.
>
> I use vtapes, and routinely leave the Amanda server and many of my
> computers running while I am away. I then SSH in and check my reports
> as email. mutt works very nicely over a slow SSH link.
>
> > Since the restart, the scheduling, which I assumed would stabilize in
> > a week or so, doesn't seemed to have settled as well as it should,
> > with the total size of the backups still varying from 14g to 35g just
> > in the last few days.
> >
> >
> >
> > Then last night, the printed report was a blank sheet of paper on the
> > output tray this morning.
>
> There I can't help you.
>
> > Fixing the scheduler would also help. It is not settling in like it
> > should. I will expand the vtape size to 50Gb for tonights run, which
> > is about 2.5x an average run needs if it is 'in balance'.
>
> I'm skeptical that that will help, but we'll see how tonight's run goes.
>
> If that doesn't help, try commenting out a bunch of DLEs, and
> uncommenting them back in slowly over the next few days. I've done that
> trick a few times in similar circumstances.
The real culprit was somewhat a comedy of errors in that that bunch of
ghostscript related updates done yesterday, required that some printers
would need to be deleted and re-installed, but the update end message
didn't say anything about that. Googling disclosed that it has been a
common occurrence with updated gs installs for the last 2-3 weeks. By
testing I found my Brother HL-2140 laser was still working, but a test page
sent to the epson used 25% of a $33 black ink tank, soaking the paper to
within a couple mm's of the edges. I tried to reconfigure it, no luck, so
I finally deleted it completely, using the cups web server at
localhost:631, then re-added it, and now a test page prints normally.
I have now added another 20Gb to the vtape size, so it should flush the
leftovers and do a normal backup tonight.
But as far as any 'balance' is concerned, I am under the impression the
scheduler is blindfolded and playing "pin the tail on the donkey" as it
doesn't seem to even be approaching a balanced use after nearly 2 more
weeks of every night runs.
[root@coyote Daily]# su amanda -c "amadmin Daily balance"
due-date #fs orig MB out MB balance
----------------------------------------------
9/22 Thu 0 0 0 ---
9/23 Fri 2 18976 10600 -18.5%
9/24 Sat 10 21077 9546 -26.6%
9/25 Sun 21 21436 13499 +3.7%
9/26 Mon 3 31429 31429 +141.5%
----------------------------------------------
TOTAL 36 92918 65074 13014
(estimated 5 runs per dumpcycle)
So that last one on the 26th will do this all over again. That sort of a
report is beginning to develop a pretty high vacuum here. I can recall,
back when my system was smaller and I was using 4Gb tapes, it could stay at
94% +- 3% media use for weeks at a time. Now, even before the extended
shutdown, I am running out of tape because it decides to do level 0's of
/usr/movies (17Gb) and /home (20 Gb) on the same nights. That sucks.
Cheers, gene
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
A word to the wise is enough.
-- Miguel de Cervantes
