Amanda-Users

Re: Resuming amanda after 2 week hiatius due to upgrading to F10.

2009-03-07 13:50:22
Subject: Re: Resuming amanda after 2 week hiatius due to upgrading to F10.
From: Gene Heskett <gene.heskett AT verizon DOT net>
To: amanda-users AT amanda DOT org
Date: Sat, 07 Mar 2009 13:45:21 -0500
On Saturday 07 March 2009, Gene Heskett wrote:
>On Saturday 07 March 2009, Gene Heskett wrote:
>>Greetings;
>>
>>I managed to destroy X somehow, and since I was running F8, it was time to
>>update to F10, so I did, but turned amanda off till things settled a bit
>> and some of the 'infant mortality' associated with an upgrade were sorted
>> I am about to restart my backup scheme, and was wondering if I should just
>> restore the line in amanda's crontab after I change the disk size in
>> amanda.conf to something on the order of 80GB so it will get all 80 some
>> GB of data in one pass, or should I leave it set at its present 15GB and
>> pre run my chatchup script which will run a backup an arbitrary number of
>> times, then enable the crontab entry once that has finished and a sort of
>> schedule established.
>>
>>Since I now have a 1TB disk for amanda to play in,, I'm inclined to try the
>>one pass gets it all, then reduce the disk size setting to something more
>>reasonable after the actual size settles some.
>>
>>This new disk is faster:
>>===============
>>[root@coyote amanda-2.6.2alpha-20090227]# hdparm -tT /dev/sdc
>>/dev/sdc:
>> Timing cached reads:   4916 MB in  2.00 seconds = 2458.87 MB/sec
>> Timing buffered disk reads:  332 MB in  3.01 seconds = 110.21 MB/sec
>>===============
>>than the /dumps disk is but not by much.  What would be the effect of de-
>>specing the holding disk and let it write directly to this bigger disk?
>>
>>And I just built the 0227 version of 262alpha and ran amccheck of course,
>> and I'll make a comment re speed:  2 years ago, amcheck usually completed
>> in some random period averaging about .9 seconds.  As 262 has progressed,
>> that is getting slower, TBE its about 2.4 seconds now, and the machine is
>> about 4-5x faster now cuz its a 2.2Ghz AMD 9550 Quad core, with 4Gb of 800
>> mhz ram to play in, where 2 years ago it was an XP-2800 single core
>> running at 1.6GHZ actual, with only a gig of 333mhz ram.
>>
>>That seems like its going backwards to me.
>>
>>Comments anyone?
>>
>>Thanks.
>
>Mmm, looks like I'm talking to myself.
>
>The first backup went well ANAICT.  It smunched the whole thing down into 44
>gigabytes and change.  And it emailed me a report as if everything worked. 
> So that part of amreport worked.  However, it is also supposed to be
> printing that report, and it did not.  There have been other occasions when
> it didn't print, but they seem to be at maybe monthly intervals, and might
> be related to something I was doing.  Using FF, whose home page I have set
> to
>localhost:6311/printers, I found a job from amanda that wasn't printed on
> Dec 31st 2008, so I emptied that queue.  Nothing else pending, and neither
> htop nor lsof can find any trace of a running amanda related function on
> the system now, so I assume it finished the amverify run also, which I am
> doing in my wrapper script.
>
>Looking up the manpage for amreport, I issued the command (as amanda)
>"amreport Daily" and got this error:
>/usr/bin/lpr: Error - no default destination available.
>amreport: printer command failed: /usr/bin/lpr
>
>But looking at the cups web page, lp1 is defined as the default printer.
>
>An lpstat -a (as amanda) returns this:
>[amanda@coyote amanda-2.6.2alpha-20090227]$ lpstat -a
>CUPS-PDF accepting requests since Thu 08 Jan 2009 01:02:22 AM EST
>lp0 accepting requests since Sun 04 Jan 2009 01:37:33 PM EST
>lp1 accepting requests since Sat 07 Mar 2009 11:57:54 AM EST
>lp2 accepting requests since Fri 28 Nov 2008 11:33:50 AM EST
>
>Which are in fact all the same printer, just different performance profiles.
>This is Fedora 10, and
>[amanda@coyote amanda-2.6.2alpha-20090227]$ ls -l /usr/bin/lpr
>lrwxrwxrwx 1 root root 23 2008-11-18 11:33 /usr/bin/lpr ->
>/etc/alternatives/print
>A link.
>
>Ok, still as amanda, I cd to /etc/alternatives and do another ls -l, which
>returns a very long list of links, and 'print' is a link to
> /usr/bin/lpr.cups.
>
>cd'ing to /usr/bin, an ls -l lpr.cups finally gets me to what should be the
>file that does the work.
>[amanda@coyote bin]$ ls -l lpr.cups
>-rwxr-xr-x 1 root root 13868 2009-01-28 13:19 lpr.cups
>
>This looks like a cups problem to me.  Permissions look ok.
>And if I try to run amreport as root, it fusses at me just like its supposed
>to.
>
>Has anyone else run into this yet?
>
>Thanks.

Further info:
logdir is "/usr/local/var/amanda/Daily"
My attempts to run amreport to get another copy are getting me emails titled:
Subject: The Coyote Den AMANDA MAIL REPORT FOR BogusMonth 0, 0

with the usual results missing for everything because:
amreport: ERROR could not open log /usr/local/var/amanda/Daily/log: No such 
file or directory

But it does exist, and this mornings logs are there.  As amanda:
ls -l /usr/local/var/amanda/Daily/log returns
-rw------- 1 amanda amanda 242040 2009-03-07 04:31 amdump.1
-rw------- 1 amanda amanda  20109 2009-03-07 04:31 log.20090307011504.0

etc etc.

So it looks like 2 problems.  Can this last one be related to this from 
ChangeLog?

009-02-24  Jean-Louis Martineau <martineau AT zmanda DOT com>
        * amandad-src/amandad.c, application-src/amgtar.c,
          changer-src/scsi-changer-driver.c, client-src/client_util.c,
          common-src/conffile.c, device-src/s3-device.c,
          oldrecover-src/set_commands.c, recover-src/set_commands.c,
          server-src/amindexd.c, server-src/planner.c,
          server-src/reporter.c: Replace all occurences of index by strchr,
                                 Replace all occurences of rindex by strrchr.


-- 
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)
<Culus> Saens demonstrates no less than 3 tcp/ip bugs in 2.2.3