Bugs item #1982210, was opened at 2008-06-02 16:35
Message generated for change (Settings changed) made by boncz
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=482468&aid=1982210&group_id=56967
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: PF/runtime
Group: Pathfinder "stable"
>Status: Closed
>Resolution: Works For Me
Priority: 5
Private: No
Submitted By: Wouter Alink (vzzzbx)
Assigned to: Nobody/Anonymous (nobody)
Summary: XQ: crash when shredding is interrupted
Initial Comment:
The following script crashes MonetDB in about 20 steps... (0.24, preJune, 28
may 2008)
QUERY='element { "test" } { for $i in 1 to 100000000 return <aap/> }'
QUERY2='pf:add-doc("/tmp/tempdoc.xml","tempdoc.xml")'
QUERY3='pf:del-doc("tempdoc.xml")'
# create a large temporary document
time echo $QUERY | mclient -lxq > /tmp/tempdoc.xml;
n=0;
while [ "$n" -ne "100" ]
do
echo "Iter $n"
# shred document
mclient -lxq -s $QUERY2 &
MAPI_PID=$!
# do not allow mclient enough time to finish shredding
sleep .01
kill $MAPI_PID
# remove document again (should fail, as it was never shredded completely)
echo $QUERY3 | mclient -lxq
n=$(($n + 1))
sleep 2
done
----------------------------------------------------------------------
Comment By: Peter Boncz (boncz)
Date: 2008-12-12 00:24
Message:
If I understand correctly, killing the Mserver resulted in a SEGV?
I tried to reproduce this in various ways, killing Mserver while
shredding, but could not reproduce this.
----------------------------------------------------------------------
Comment By: Wouter Alink (vzzzbx)
Date: 2008-06-04 15:37
Message:
Logged In: YES
user_id=621590
Originator: YES
stefan, i wasn't trying it at home ;)
i just needed some document, for which it was guaranteed that the kill
signal would arrive earlier than that the shred-doc completed, and it
seemed that this one would do the trick :)
----------------------------------------------------------------------
Comment By: Stefan Manegold (stmane)
Date: 2008-06-03 18:56
Message:
Logged In: YES
user_id=572415
Originator: NO
File Added: server.10000000.log
----------------------------------------------------------------------
Comment By: Stefan Manegold (stmane)
Date: 2008-06-03 18:56
Message:
Logged In: YES
user_id=572415
Originator: NO
File Added: client.10000000.log
----------------------------------------------------------------------
Comment By: Stefan Manegold (stmane)
Date: 2008-06-03 18:56
Message:
Logged In: YES
user_id=572415
Originator: NO
File Added: server.1000000.log
----------------------------------------------------------------------
Comment By: Stefan Manegold (stmane)
Date: 2008-06-03 18:56
Message:
Logged In: YES
user_id=572415
Originator: NO
Ok, in my 8 GB Fedora 8 desktop, Mserver survives the script with
documents of 100000 & 1000000 nodes (0.58 MB & 5.8 MB),
but indeed crashes with a
*** glibc detected *** Mserver: double free or corruption (fasttop):
0x0000000001208e90 ***
after 17 Iterations with a document of 10000000 nodes (58 MB).
See attached logs for details.
Further analysis pending.
Any help, ideas, suggestions are more than welcome!
File Added: client.1000000.log
----------------------------------------------------------------------
Comment By: Stefan Manegold (stmane)
Date: 2008-06-03 18:39
Message:
Logged In: YES
user_id=572415
Originator: NO
How does Mserver crash, i.e., is there any message, and if so, which one?
(Such info always helps analysis & debugging a lot ...)
----------------------------------------------------------------------
Comment By: Stefan Manegold (stmane)
Date: 2008-06-03 18:29
Message:
Logged In: YES
user_id=572415
Originator: NO
WARNING:
DO *NOT* TRY THIS AT HOME!
Already the first query to "create a large temporary document" will make
Mserver grow to >> 20 GB !
Is indeed such a huge document required to reproduce the problem???
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=482468&aid=1982210&group_id=56967
------------------------------------------------------------------------------
SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada.
The future of the web can't happen without you. Join us at MIX09 to help
pave the way to the Next Web now. Learn more and register at
http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/
_______________________________________________
Monetdb-bugs mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/monetdb-bugs
