This is a public chat log generated from the #semsol IRC channel.
02:14:03
i'm trying to get arc2 to run and having a rather bad time of it
02:15:58
Fatal error: Uncaught exception 'ErrorException' with message 'Non-static method ARC2::getStore() should not be called statically' in /home/wholeithaca/sites/animalphotos.info/apps/flickr_pipeline/bin/create_infobox.php:14
02:18:36
i look at the code and I don't see anything is static in ARC2.php, version 2008-09-30
02:18:58
the keyword static doesn't even show up in the code
02:20:30
so I make an instance of ARC2.php and try calling getStore() as an instance method and I get
02:20:48
Fatal error: Uncaught exception 'ErrorException' with message 'Redefining already defined constructor for class ARC2_Store' in /home/wholeithaca/sites/animalphotos.info/apps/flickr_pipeline/arc/store/ARC2_Store.php:19
02:21:25
What's going on here?
02:28:35
is the problem that I'm running PHP 5?
02:31:50
i see what's going on... I've got strict error checking on... That's a bummer
02:59:40
Seems to work OK if E_STRICT is off...
18:51:07
i've been having fun loading dbpedia's turtle files in arc... loading about 32M tuples... started out around 1000 tuples/sec, now with 3M tuples in, it's doing about 330
18:53:25
found one hell of a bug in the turtle parser around 1 am last nice... did a grep -v of a 4G file to remove about 30 offending lines, now it's cooking
18:55:17
tried seeing if i could improve performance by trying different i/o schedulers... nothing helped much, but picking the wrong scheduler made the web server on the machine slow down to a crawl
18:57:30
the turtle parser doesn't like "strings that end with \\"
19:01:30
if storage scales linearly, the myisamdatabase is going to be 7-8 GB... Not too bad, particularly after I run it through the table compressor
19:04:36
the machine has a pretty crap RAID away if you ask me... if it brings in some more cash, I might add a non-RAID SAS drive for doing this sort of thing
19:47:37
does ARC use the SPARQL Query Results XML Format?
19:56:44
for input or output?
19:58:39
output
20:01:25
http://arc.semsol.org/docs/v2/endpoint
20:01:32
^---- i think so
20:02:09
i'll know for sure when i get my 30 M tuples loaded
20:02:37
they'll be some trouble if a parser error kills it halfway through
20:04:07
can't complain about parser speed... my i/o system is pegged out at almost 100%, almost all writes
20:09:24
at 5M tuples I'm probably averaging 160 tuples/sec now. very bursty. i've probably got the only old fogey RAID controller that really flushes when you call flush()
20:43:45
thx