Locking problem

19 views
Skip to first unread message

sastry

unread,
Oct 8, 2009, 6:02:24 AM10/8/09
to jBASE
Hi,

Here we have a strange situation when stale sessions like when use
abruptly closes the session or session closed abruplty without any
user intervention then the lock acquiered by this processed is not
getting released approx 13 to 14 minutes where we have set jRLA tidyup
time to 1minute .Any clues??

Best Regards,
Sastry

sastry

unread,
Oct 9, 2009, 6:43:53 AM10/9/09
to jBASE
Find the jRLA output from the system

jsh test ~ -->SHOW-RECORD-LOCKS
Polling 4 ports ...
PORT PID
FILENAME
RECORDKEY LOCK#
PORT/-PID

1 413928
UNKNOWN
432105163280008
0x0db8e5e4,W ---
1 413928
UNKNOWN
FT09279007294309
0x1cf126b8,W ---
1 413928
UNKNOWN
425105358590004
0x773daeac,W ---
2 405654
UNKNOWN
425101264830006
0x0d1c578d,W ---
2 405654
UNKNOWN
FT09279002007302
0x125de6cc,W ---
2 405654 ../bnk.data/ac/
FBNK.ACCOUNT
425105358590004
00000000,W 1
3 139716
UNKNOWN
FT09279007610720
0x64df432d,W ---
3 139716
UNKNOWN
423101027780005
0x2d6fd6eb,W ---
3 139716 ../bnk.data/ac/
FBNK.ACCOUNT
425105358590004
00000000,W 1
jsh test ~ -->ls -ltr ../bnk.data/ac/FBNK.ACCOUNT
-rwxrwxrwx 1 r8env1 staff 918929408 Oct 8 14:06 ../bnk.data/
ac/FBNK.ACCOUNT
jsh r8env1 ~ -->jRLA -dvL
Program at pid: 168300
Program started: Thu Oct 8 14:45:16 2009
Time last ran: Fri Oct 9 13:36:22 2009
Time last tidied up: Fri Oct 9 13:35:24 2009
Lock retries: 2123
Tidy-up time: 1 minute
Locking mechanism: IPC semaphores based on THREAD id
Tidy-up operations: 0
Semaphore IDs: 46137356, 25165835
Memory: 525920 bytes used , 123104 bytes free
CPU Usage: 3.98 usr , 4.56 sys

Record locks outstanding
Group value pid type i-node device queued key
48 0x773daeac 413928 WRITE 0020ad0c 00000002 None
425105358590004
189 0x2d6fd6eb 139716 WRITE 0020ad0c 00000002 None
423101027780005
240 0x125de6cc 405654 WRITE 002db832 00000002 None
FT09279002007302
618 0x64df432d 139716 WRITE 002db832 00000002 None
FT09279007610720
803 0x1cf126b8 413928 WRITE 002db832 00000002 None
FT09279007294309
847 0x0d1c578d 405654 WRITE 0020ad0c 00000002 None
425101264830006
924 0x0db8e5e4 413928 WRITE 0020ad0c 00000002 None
432105163280008

Record locks: 30930 locks maximum in 1031 groups of 30 locks/
group
7 locks currently in use
1 is max. locks found in one group

Even after setting the tidyup time to 1 minutes the stale sessions
remains locks after 1 minute

Following is the jdiag output


JBCPORTNO : '1-2500'
JBCRELEASEDIR : '/usr/jbc5017'
JBCGLOBALDIR : '/usr/jbc5017'
WARNING: JBCDATADIR is not set, Default '/usr/jbc5017/jbase_data'
WARNING: JBCDATADIR is subdirectory of JBCGLOBALDIR
HOME : '/globus/test/bnk/bnk.run'
JEDIFILEPATH : '/globus/test/bnk/bnk.run:/globusprod/
bab/bnk/bnk.data/dst'
JEDIFILENAME_MD : '/globus/test/bnk/bnk.run/VOC'
JEDIFILENAME_SYSTEM : '/usr/jbc5017/src/SYSTEM'
RELEASE Information : Major 5.0 , Minor 17 , Patch 0335
(Change 80765)
Spooler dir (JBCSPOOLERDIR) : '/globus/test/bnk/bnk.run/jspooler'
Spooler directory '/globus/test/bnk/bnk.run/jspooler' OK
JBCEMULATE : 'prime'
Emulation Config file '/usr/jbc5017/config/Config_EMULATE' OK
JBCEMULATE Label 'prime' found in file '/usr/jbc5017/config/
Config_EMULATE'

pat

unread,
Oct 9, 2009, 5:27:02 PM10/9/09
to jBASE
What does :

WHERE (V

show for Ports 1, 2 and 3 ?

And :

ps -ef

show for 'pids' 413928, 405654 and 139716 ?

Pat.

sastry

unread,
Oct 10, 2009, 3:05:37 AM10/10/09
to jBASE
HI Pat,
Thanks for your reply,WHERE output for the port 1 was showing
I_SF.INPUT as user abruptly closed when waiting for a input on
override. other two ports are spending in LOCK block.

Best Regards,
Sastry
> > > Sastry- Hide quoted text -
>
> - Show quoted text -

Jim Idle

unread,
Oct 16, 2009, 3:24:44 AM10/16/09
to jb...@googlegroups.com
I would suggest that when the user closes the session the process they wre attached to does not actually disappear for 13 to 14 minutes. When it finally does, then the locks will be cleaned up. You should check these processes to see if they are spinning like mad eating CPU trying to read input etc.

Jim

Tec Murho

unread,
Oct 26, 2009, 3:36:50 AM10/26/09
to jb...@googlegroups.com
you can boost your system by automating this !!
trick things to build a scripts to kill the hanged ps,
and check also your perf,

Jim Idle

unread,
Oct 26, 2009, 12:52:17 PM10/26/09
to jb...@googlegroups.com

> -----Original Message-----
> From: jb...@googlegroups.com [mailto:jb...@googlegroups.com] On Behalf
> Of Tec Murho
> Sent: Monday, October 26, 2009 12:37 AM
> To: jb...@googlegroups.com
> Subject: Re: Locking problem
>
>
> you can boost your system by automating this !!
> trick things to build a scripts to kill the hanged ps,
> and check also your perf,


For all the many reasons I have been repeating for many years, do NOT write scripts that continually run on the system looking for 'problems' and definitely do not write automated kill scripts.

Jim

Daniel Klein

unread,
Oct 26, 2009, 12:59:06 PM10/26/09
to jb...@googlegroups.com
Maybe you should write a script to continually repeat this email ;-)

Dan

Jim Idle

unread,
Oct 26, 2009, 1:03:01 PM10/26/09
to jb...@googlegroups.com

> -----Original Message-----
> From: jb...@googlegroups.com [mailto:jb...@googlegroups.com] On Behalf
> Of Daniel Klein
> Sent: Monday, October 26, 2009 9:59 AM
> To: jb...@googlegroups.com
> Subject: Re: Locking problem
>
>

> Maybe you should write a script to continually repeat this email ;-)


You're fired!

Jim


Jim Idle

unread,
Oct 26, 2009, 1:05:24 PM10/26/09
to jb...@googlegroups.com

> -----Original Message-----
> From: jb...@googlegroups.com [mailto:jb...@googlegroups.com] On Behalf
> Of Daniel Klein
> Sent: Monday, October 26, 2009 9:59 AM
> To: jb...@googlegroups.com
> Subject: Re: Locking problem
>
>

> Maybe you should write a script to continually repeat this email ;-)
>

You're fired.

:: Automated Firing Script version 5.2.1 - Loop detected at "Daniel Klein"... aborting in case bug fires others by accident... ::


Jim Idle

unread,
Oct 26, 2009, 1:06:18 PM10/26/09
to jb...@googlegroups.com

> -----Original Message-----
> From: jb...@googlegroups.com [mailto:jb...@googlegroups.com] On Behalf
> Of Daniel Klein
> Sent: Monday, October 26, 2009 9:59 AM
> To: jb...@googlegroups.com
> Subject: Re: Locking problem
>
>

> Maybe you should write a script to continually repeat this email ;-)
>
> Dan

You're fired.

:: Automated Firing Script version 5.2.1 - Unable to abort, system issue, monitoring script exhausted all resources... ::


Pawel (privately)

unread,
Oct 27, 2009, 2:17:45 PM10/27/09
to jb...@googlegroups.com
Dnia 26-10-2009 o godz. 18:06 Jim Idle napisał(a):
> > Maybe you should write a script to continually repeat this email ;-)
> >
> > Dan
>
> You're fired.
>
> :: Automated Firing Script version 5.2.1 - Unable to abort, system
> issue, monitoring script exhausted all resources... ::
I personally advise to write additional monitoring script to monitor the
one written by Jim ;) Definetely something does not work in Jim's
script, because it failed with "out of resources" message. <g>

Kind regards
Pawel

----------------------------------------------------
Film Solista: 9/10 Jak dla mnie rewelacyjny!
Piękny wzruszający głęboki!
Trudna tematyka, ale pokazana w mistrzowski sposób.
http://klik.wp.pl/?adr=http%3A%2F%2Fcorto.www.wp.pl%2Fas%2Fsolista.html&sid=900


Jim Idle

unread,
Oct 27, 2009, 4:01:13 PM10/27/09
to jb...@googlegroups.com
And of course, that was the subtle point ;-)

Jim

> -----Original Message-----
> From: jb...@googlegroups.com [mailto:jb...@googlegroups.com] On Behalf
Reply all
Reply to author
Forward
0 new messages