[MonetDB-users] MServer crashing: Random Segmentation fault
Hi all, I am using MD4 with XQuery + PF/Tijah module. I have a problem that my server sometimes/frequently crashes. I could not detect a trigger, as it occurs randomly and at any time. It gives me this error: 26747 Segmentation fault Mserver --dbinit="module(pathfinder);" but it gives me any number before it. I am not sure where I can locate the MD error log (if it exists), and how it is triggered. Help? Thanks, junte
Hi junte,
can you please tell which version you are using? stable release, night
builds or checkout from cvs?
Also, do you use mclient to pose queries? do you use the -G flag with mclient?
It will be also useful to give us one example of the sequence of
queries you use so we can pinpoint the problem.
thank you,
lefteris
On Fri, Apr 18, 2008 at 12:51 PM, jz@uva
Hi all,
I am using MD4 with XQuery + PF/Tijah module.
I have a problem that my server sometimes/frequently crashes. I could not detect a trigger, as it occurs randomly and at any time.
It gives me this error:
26747 Segmentation fault Mserver --dbinit="module(pathfinder);"
but it gives me any number before it.
I am not sure where I can locate the MD error log (if it exists), and how it is triggered.
Help?
Thanks, junte
------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javao... _______________________________________________ MonetDB-users mailing list MonetDB-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-users
Hi lefteris,
On Fri, Apr 18, 2008 at 1:16 PM, Lefteris
Hi junte,
can you please tell which version you are using? stable release, night builds or checkout from cvs?
I am using the stable release from February. I compiled this release from CVS on Fedora Core 8.
Also, do you use mclient to pose queries? do you use the -G flag with mclient?
Yes, I am using mclient for posing queries. I am not using the -G flag for mclient. It is the algebra frontend. Do I need to set this flag? My XQueries are standard, and they do work, so I think that may not be the problem. Thanks, junte
It will be also useful to give us one example of the sequence of queries you use so we can pinpoint the problem.
thank you,
lefteris
On Fri, Apr 18, 2008 at 12:51 PM, jz@uva
wrote: Hi all,
I am using MD4 with XQuery + PF/Tijah module.
I have a problem that my server sometimes/frequently crashes. I could not detect a trigger, as it occurs randomly and at any time.
It gives me this error:
26747 Segmentation fault Mserver --dbinit="module(pathfinder);"
but it gives me any number before it.
I am not sure where I can locate the MD error log (if it exists), and how it is triggered.
Help?
Thanks, junte
------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javao... _______________________________________________ MonetDB-users mailing list MonetDB-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-users
------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javao... _______________________________________________ MonetDB-users mailing list MonetDB-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-users
On 04/18/2008 12:51 PM, jz@uva wrote with possible deletions:
Hi all,
I am using MD4 with XQuery + PF/Tijah module.
I have a problem that my server sometimes/frequently crashes. I could not detect a trigger, as it occurs randomly and at any time.
It gives me this error:
26747 Segmentation fault Mserver --dbinit="module(pathfinder);"
but it gives me any number before it.
I am not sure where I can locate the MD error log (if it exists), and how it is triggered.
You could give us a hint by starting the Mserver in a debugging session: $ gdb --args Mserver --dbinit='module(pathfinder); mil_start();' start the Mserver with: run and create a dump as soon as you triggered the segmentation fault with: bt This should give us an idea where your problem is located. Jan -- Jan Rittinger Database Systems Technische Universität München (Germany) http://www-db.in.tum.de/~rittinge/
On Fri, Apr 18, 2008 at 3:41 PM, Jan Rittinger
You could give us a hint by starting the Mserver in a debugging session:
$ gdb --args Mserver --dbinit='module(pathfinder); mil_start();'
start the Mserver with: run and create a dump as soon as you triggered the segmentation fault with: bt
This should give us an idea where your problem is located.
Thanks, this is the error I am getting: MonetDB>[New Thread -1319306352 (LWP 18169)] Program received signal SIGSEGV, Segmentation fault. [Switching to Thread -1319306352 (LWP 18169)] 0x0047c201 in ll_upwards () from /home/xxx/Development/pftijah/lib/MonetDB4/lib/lib_pf_support.so Missing separate debuginfos, use: debuginfo-install bzip2.i386 e2fsprogs.i386 glibc.i686 keyutils.i386 krb5.i386 libselinux.i386 openssl.i686 zlib.i386 I hope this helps.... junte
------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javao... _______________________________________________ MonetDB-users mailing list MonetDB-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-users
I forgot, output from bt:
(gdb) bt
#0 0x0047c201 in ll_upwards () from
/home/xxx/Development/pftijah/lib/MonetDB4/lib/lib_pf_support.so
#1 0x011e0006 in seqscan_LH_sht_void_tloc_var_var () from
/home/xxx/Development/pftijah/lib/libbat.so.0
#2 0x00000000 in ?? ()
junte
On Fri, Apr 18, 2008 at 4:21 PM, jz@uva
On Fri, Apr 18, 2008 at 3:41 PM, Jan Rittinger
wrote: You could give us a hint by starting the Mserver in a debugging session:
$ gdb --args Mserver --dbinit='module(pathfinder); mil_start();'
start the Mserver with: run and create a dump as soon as you triggered the segmentation fault with: bt
This should give us an idea where your problem is located.
Thanks, this is the error I am getting:
MonetDB>[New Thread -1319306352 (LWP 18169)]
Program received signal SIGSEGV, Segmentation fault. [Switching to Thread -1319306352 (LWP 18169)] 0x0047c201 in ll_upwards () from /home/xxx/Development/pftijah/lib/MonetDB4/lib/lib_pf_support.so Missing separate debuginfos, use: debuginfo-install bzip2.i386 e2fsprogs.i386 glibc.i686 keyutils.i386 krb5.i386 libselinux.i386 openssl.i686 zlib.i386
I hope this helps....
junte
------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javao... _______________________________________________ MonetDB-users mailing list MonetDB-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-users
Could you update your CVS version of pathfinder and try again? If I remember correctly there were some post-release bug fixes to the upward step. (If updating doesn't help could you please report a bug -- if possible with a simplified example query.) Cheers, Jan On 04/18/2008 04:23 PM, jz@uva wrote with possible deletions:
I forgot, output from bt:
(gdb) bt #0 0x0047c201 in ll_upwards () from /home/xxx/Development/pftijah/lib/MonetDB4/lib/lib_pf_support.so #1 0x011e0006 in seqscan_LH_sht_void_tloc_var_var () from /home/xxx/Development/pftijah/lib/libbat.so.0 #2 0x00000000 in ?? ()
junte
On Fri, Apr 18, 2008 at 4:21 PM, jz@uva
wrote: On Fri, Apr 18, 2008 at 3:41 PM, Jan Rittinger
wrote: You could give us a hint by starting the Mserver in a debugging session:
$ gdb --args Mserver --dbinit='module(pathfinder); mil_start();'
start the Mserver with: run and create a dump as soon as you triggered the segmentation fault with: bt
This should give us an idea where your problem is located.
Thanks, this is the error I am getting:
MonetDB>[New Thread -1319306352 (LWP 18169)]
Program received signal SIGSEGV, Segmentation fault. [Switching to Thread -1319306352 (LWP 18169)] 0x0047c201 in ll_upwards () from /home/xxx/Development/pftijah/lib/MonetDB4/lib/lib_pf_support.so Missing separate debuginfos, use: debuginfo-install bzip2.i386 e2fsprogs.i386 glibc.i686 keyutils.i386 krb5.i386 libselinux.i386 openssl.i686 zlib.i386
I hope this helps....
junte
------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javao... _______________________________________________ MonetDB-users mailing list MonetDB-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-users
------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javao... _______________________________________________ MonetDB-users mailing list MonetDB-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-users
-- Jan Rittinger Database Systems Technische Universität München (Germany) http://www-db.in.tum.de/~rittinge/
On Fri, Apr 18, 2008 at 5:37 PM, Jan Rittinger
Could you update your CVS version of pathfinder and try again?
If I remember correctly there were some post-release bug fixes to the upward step.
(If updating doesn't help could you please report a bug -- if possible with a simplified example query.)
Cheers, Jan
Hi Jan, Sorry for my late reply. I have installed the latest stable February release CVS version a week ago. Is it useful to reinstall again now? If not, I will submit a bug report. Thanks, Junte
On 04/18/2008 04:23 PM, jz@uva wrote with possible deletions:
I forgot, output from bt:
(gdb) bt #0 0x0047c201 in ll_upwards () from /home/xxx/Development/pftijah/lib/MonetDB4/lib/lib_pf_support.so #1 0x011e0006 in seqscan_LH_sht_void_tloc_var_var () from /home/xxx/Development/pftijah/lib/libbat.so.0 #2 0x00000000 in ?? ()
junte
On Fri, Apr 18, 2008 at 4:21 PM, jz@uva
wrote: On Fri, Apr 18, 2008 at 3:41 PM, Jan Rittinger
wrote: You could give us a hint by starting the Mserver in a debugging session:
$ gdb --args Mserver --dbinit='module(pathfinder); mil_start();'
start the Mserver with: run and create a dump as soon as you triggered the segmentation fault with: bt
This should give us an idea where your problem is located.
Thanks, this is the error I am getting:
MonetDB>[New Thread -1319306352 (LWP 18169)]
Program received signal SIGSEGV, Segmentation fault. [Switching to Thread -1319306352 (LWP 18169)] 0x0047c201 in ll_upwards () from /home/xxx/Development/pftijah/lib/MonetDB4/lib/lib_pf_support.so Missing separate debuginfos, use: debuginfo-install bzip2.i386 e2fsprogs.i386 glibc.i686 keyutils.i386 krb5.i386 libselinux.i386 openssl.i686 zlib.i386
I hope this helps....
junte
------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javao... _______________________________________________ MonetDB-users mailing list MonetDB-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-users
------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javao... _______________________________________________ MonetDB-users mailing list MonetDB-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-users
-- Jan Rittinger Database Systems Technische Universität München (Germany) http://www-db.in.tum.de/~rittinge/
-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javao... _______________________________________________ MonetDB-users mailing list MonetDB-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-users
On 04/24/2008 01:20 PM, jz@uva wrote with possible deletions:
On Fri, Apr 18, 2008 at 5:37 PM, Jan Rittinger
wrote: Could you update your CVS version of pathfinder and try again?
If I remember correctly there were some post-release bug fixes to the upward step.
(If updating doesn't help could you please report a bug -- if possible with a simplified example query.)
Cheers, Jan
Hi Jan,
Sorry for my late reply. I have installed the latest stable February release CVS version a week ago. Is it useful to reinstall again now?
Did you get the latest version with bugfixes using 'cvs up -dP' in the pathfinder folder? -- If not I would propose to update pathfinder, re-compile pathfinder, and to test again.
If not, I will submit a bug report.
If my above proposal does not change the behavior I would ask you to report the bug. -- Jan Rittinger Database Systems Technische Universität München (Germany) http://www-db.in.tum.de/~rittinge/
On 24-04-2008 13:30:29 +0200, Jan Rittinger wrote:
Sorry for my late reply. I have installed the latest stable February release CVS version a week ago. Is it useful to reinstall again now?
Did you get the latest version with bugfixes using 'cvs up -dP' in the pathfinder folder? -- If not I would propose to update pathfinder, re-compile pathfinder, and to test again.
If not, I will submit a bug report.
If my above proposal does not change the behavior I would ask you to report the bug.
./monetdb-install.sh --cvs --enable-xquery
participants (4)
-
Fabian Groffen
-
Jan Rittinger
-
jz@uva
-
Lefteris