Re: etimeout ignored?
2007-09-04 11:10:39
> Paul Bijnens schrieb:
>> How did you came to the conclusion that the server still waits for the
>> estimates? amstatus? which output? of just the fact that amanda is
>> still running? or...
Ralf Auer <Ralf.Auer AT physik.uni-erlangen DOT de> writes:
> a 'amgetconf NewSetup etimeout' returns '900'. So, it seems to be
> configured correctly.
Interesting, I'm running 2.5.1p1-2.1 from Debian packages, and I get:
$ amgetconf offsite etimeout
amgetconf: getconf_str: np is not a CONFTYPE_STRING|CONFTYPE_IDENT: 26
I'm not entirely sure what this means. However, I'm certainly taking
*this* as something bad:
$ amadmin offsite config| grep -i etimeout
ETIMEOUT 220000
$ grep etimeout offsite/amanda.conf
etimeout 10800 # number of seconds per filesystem for estimates.
So, for some reason, amanda is running with an expectation of 61.11
hours of timeout *per* DLE? No wonder this is taking forever!
> I came to the conclusion that my server still waits for the estimate
> by issuing the 'amstatus' command. It told me, that the server is
> still 'waiting for estimate' for one host, all other hosts were in
> 'estimate done' state.
Which is exactly what I did. I also looked at the running processes
(like planner) and straced them to verify they were in a holding
pattern.
> And last night I watched my backup running
I'm still watching mine, it's much like watching the grass grow.
Except, the lawn makes progress and requires me to do something
occassionally :)
>> Another frequent "monday-morning-no-coffee-yet" problem encountered is
>> that you're looking at the wrong config file, or etimeout appears twice
>> in the config file. Verify with:
>>
>> amgetconf daily etimeout
>>
>> You can also set etimeout to a negative value, to avoid the
>> multiplication of the number of DLE's by the etimeout value.
I'll try the negative value, but I'm *very* curious to know where
amanda got an etimeout value of 220000? I've checked my other
timeouts as well, and they're also completely wrong:
$ grep -i timeout offsite/amanda.conf
etimeout 10800 # number of seconds per filesystem for estimates.
dtimeout 1800 # number of idle seconds before a dump is aborted.
ctimeout 30 # maximum number of seconds that amcheck waits
$ amadmin offsite config| grep -i timeout
ETIMEOUT 220000
DTIMEOUT 193000
CTIMEOUT 190030
I just did a quick test and set my etimeout in the amanda.conf file to
30. I then did:
$ amadmin offsite config| grep -i timeout
ETIMEOUT 190030
DTIMEOUT 193000
CTIMEOUT 190030
Why is amanda adding to the timeout? And by what algorithm ? It adds
190000 for thet e/ctimeouts, but 191200 for the dtimeout? Yet when
etimeout is set to 10800, it adds 209200 ? These numbers don't seem
like unix time values, either...
Something seems very broken...
I'll be glad to add my entire amanda.conf setup if anyone thinks it's useful.
--
Thanks,
Paul
|
|
|