I had a job that sent it’s “I’m Finished” email with a timestamp of 0208,
but it was probably only doing tape work between 02:00 and the new 02:00.
Yeah, the last dumper activity seems to have finished at 01:26 — the first
01:26 —
so only same-node taper work was going on.
Looks like I got lucky:
Lets see:
taper.20141101080003.debug shows the time jumping back suddenly but nobody
seems to care:
Sun Nov 2 01:54:14 2014: thd-0x8b96ff0: taper: Amanda::Taper::Scribe preparing
to write, part size 0, using no cache (PEOM will be fatal) (splitter) (no LEOM)
Sun Nov 2 01:54:14 2014: thd-0x8b96ff0: taper: Starting <Xfer@0x96d08c8
(<XferSourceHolding@0x9345938> -> <XferDestTaperSplitter@0x93451e0>)>
Sun Nov 2 01:54:14 2014: thd-0x8b96ff0: taper: Final linkage:
<XferSourceHolding@0x9345938> -(PULL_BUFFER)-> <XferElementGlue@0x934d150>
-(PUSH_BUFFER)-> <XferDestTaperSplitter@0x93451e0>
Sun Nov 2 01:54:14 2014: thd-0x9827c80: taper: Building type SPLIT_FILE header
of 32768-32768 bytes with name='clxsrv.fnal.gov' disk='/export/userb'
dumplevel=0 and blocksize=32768
Sun Nov 2 01:00:09 2014: thd-0x8b96ff0: taper: Amanda::Taper::Scribe preparing
to write, part size 0, using no cache (PEOM will be fatal) (splitter) (no LEOM)
Sun Nov 2 01:00:09 2014: thd-0x8b96ff0: taper: Starting <Xfer@0x996c990
(<XferSourceHolding@0x93458c8> -> <XferDestTaperSplitter@0x9345108>)>
Sun Nov 2 01:00:09 2014: thd-0x8b96ff0: taper: Final linkage:
<XferSourceHolding@0x93458c8> -(PULL_BUFFER)-> <XferElementGlue@0x934d0a8>
-(PUSH_BUFFER)-> <XferDestTaperSplitter@0x9345108>
Sun Nov 2 01:00:09 2014: thd-0x99d8da8: taper: Building type SPLIT_FILE header
of 32768-32768 bytes with name='daesrv.fnal.gov' disk='/export/engines'
dumplevel=0 and blocksize=32768
Sun Nov 2 01:04:30 2014: thd-0x8b96ff0: taper: Amanda::Taper::Scribe preparing
to write, part size 0, using no cache (PEOM will be fatal) (splitter) (no LEOM)
dumper.20141101080002007.debug seems to be temporarily not busy, but did
get a QUIT command at 2:07
(if I got the AMREPORT at 02:08, this is when every piece was wrapping up
and being told to QUIT)
It looks like he was done transferring data at the first 01:26, remained
totally ignorant of the next hour having
repeated timestamps, cuz he wasn’t doing anything, and then wrote down
02:07 for the QUIT after the time had changed
There was no wrapping of the timestamps, so I think there is a whole 90
minutes actually between that 01:26 entry and the 02:07 one.
Sun Nov 2 01:26:18 2014: thd-0x8736368: dumper: stream_read_callback: data is
still flowing
Sun Nov 2 01:26:29 2014: thd-0x8736368: dumper: stream_read_callback: data is
still flowing
Sun Nov 2 01:26:39 2014: thd-0x8736368: dumper:
security_stream_seterr(0x87f50d0, EOF)
Sun Nov 2 01:26:39 2014: thd-0x8736368: dumper:
security_stream_close(0x87f50d0)
Sun Nov 2 01:26:39 2014: thd-0x8736368: dumper:
security_stream_seterr(0x87e5060, EOF)
Sun Nov 2 01:26:39 2014: thd-0x8736368: dumper:
security_stream_close(0x87e5060)
Sun Nov 2 01:26:39 2014: thd-0x8736368: dumper:
security_stream_seterr(0x87ed098, EOF)
Sun Nov 2 01:26:39 2014: thd-0x8736368: dumper:
security_stream_close(0x87ed098)
Sun Nov 2 01:26:39 2014: thd-0x8736368: dumper: Building type FILE header of
32768-32768 bytes with name='clxsrv.fnal.gov' disk='/mecca_head' dumplevel=0
and blocksize=32768
Sun Nov 2 01:26:39 2014: thd-0x8736368: dumper: putresult: 3 DONE
Sun Nov 2 02:07:28 2014: thd-0x8736368: dumper: getcmd: QUIT ""
Sun Nov 2 02:07:28 2014: thd-0x8736368: dumper: pid 3135 finish time Sun Nov
2 02:07:28 2014
On Nov 3, 2014, at 8:52 AM, Gene Heskett <[email protected]> wrote:
> Greetings;
>
> Running 3.3.6 here.
>
> Last night it played catchup ok, but when DST ended at 2am Sunday morning,
> apparently all the timeout timers in amanda went aglay, and it tried to
> quit, but left numerous bits and pieces still running, hogging the machine
> heavily. It sent me the email as if it was finished and had failed most
> DLE's. Half an hour later there were still copies of chunker and gzip
> using 99% of the cpu. I did eventually find the parent that was
> restarting stuff every time I killed it, but eventually I gave up and
> rebooted as I was still having 1 minute+ pauses between pressing a key and
> seeing it on screen.
>
> Now that I think about it, ISTR I wound up rebooting it when the time went
> fwd last spring.
>
> I know this isn't much to go on, but I thought it ought to be reported
> that amanda gets all flustered when the clocks are changed while it is
> running.
>
> Did anyone else have an amdump session running at 2 am Sunday? Did it
> work, or blow up?
>
> 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
