Michael,
I've had a few goes at trying to reproduce this without success
(developmenet version - if you used one of the loaders, then I can't
think of anything post build 27 (20120829) that might play into this,
but that is not to say there isn't anything).
What I don't understand is where the garbage is coming from.
It may be the queries, and not the update. While the design is MVCC
(writers not blocked by readers), that is not to say that under load
there isn't a starvation situation where the writer is implicitly held
up but again the code in fact slightly favours writers.
So does the log provide any clues? (Running with -v provides more
details - including the updates).
Andy
On 13/09/12 18:04, Michael Brunnbauer wrote:
Hello Andy,
On Thu, Sep 13, 2012 at 05:46:50PM +0100, Andy Seaborne wrote:
What is the update? is it just a
DROP SILENT GRAPH <http://voordijk.eu/foaf.axd>
yes.
and if so, how big is the graph <http://voordijk.eu/foaf.axd>?
It's not there (I wrote that already).
Is journal.jrnl of non-zero size?
no - and the timestamp is way before the update.
nodes.dat-jrnl and prefixes.dat-jrnl also have 0 size but current timestamps:
-rw-r--r-- 1 java java 38419824640 Sep 2 13:29 GOSP.dat
-rw-r--r-- 1 java java 176160768 Sep 2 13:29 GOSP.idn
-rw-r--r-- 1 java java 38419824640 Sep 2 07:24 GPOS.dat
-rw-r--r-- 1 java java 176160768 Sep 2 07:24 GPOS.idn
-rw-r--r-- 1 java java 38419824640 Sep 1 23:55 GSPO.dat
-rw-r--r-- 1 java java 176160768 Sep 1 23:55 GSPO.idn
-rw-r--r-- 1 java java 8388608 Aug 31 10:28 OSP.dat
-rw-r--r-- 1 java java 8388608 Aug 31 10:28 OSP.idn
-rw-r--r-- 1 java java 38419824640 Sep 3 12:52 OSPG.dat
-rw-r--r-- 1 java java 176160768 Sep 3 12:52 OSPG.idn
-rw-r--r-- 1 java java 8388608 Aug 31 10:28 POS.dat
-rw-r--r-- 1 java java 8388608 Aug 31 10:28 POS.idn
-rw-r--r-- 1 java java 38419824640 Sep 3 06:47 POSG.dat
-rw-r--r-- 1 java java 176160768 Sep 3 06:47 POSG.idn
-rw-r--r-- 1 java java 8388608 Aug 31 10:28 SPO.dat
-rw-r--r-- 1 java java 8388608 Aug 31 10:28 SPO.idn
-rw-r--r-- 1 java java 38419824640 Sep 2 19:16 SPOG.dat
-rw-r--r-- 1 java java 176160768 Sep 2 19:16 SPOG.idn
-rw-r--r-- 1 java java 0 Sep 3 13:21 fixed.opt
-rw-r--r-- 1 java java 0 Sep 3 13:38 journal.jrnl
-rw-r--r-- 1 java java 10594811904 Sep 1 17:25 node2id.dat
-rw-r--r-- 1 java java 67108864 Sep 1 17:25 node2id.idn
-rw-r--r-- 1 java java 24941517653 Sep 1 18:03 nodes.dat
-rw-r--r-- 1 java java 0 Sep 13 13:54 nodes.dat-jrnl
-rw-r--r-- 1 java java 8388608 Aug 31 10:28 prefix2id.dat
-rw-r--r-- 1 java java 8388608 Aug 31 10:28 prefix2id.idn
-rw-r--r-- 1 java java 83what's in teh log 88608 Aug 31 10:28 prefixIdx.dat
-rw-r--r-- 1 java java 8388608 Aug 31 10:28 prefixIdx.idn
-rw-r--r-- 1 java java 0 Aug 31 10:28 prefixes.dat
-rw-r--r-- 1 java java 0 Sep 13 13:54 prefixes.dat-jrnl
-rw-r--r-- 1 java java 2761103 Sep 1 18:08 stats.opt.old
Regards,
Michael Brunnbauer
