similar to: How to detect Deadlock using DTrace

Hi, I have an JAVA application using JNI calls which links to native libraries.When i run this application for around 4hrs i am getting a java.io.IOException "Not enough space". No memory leaks in JAVA as my JProbe reports are fine. I came across DTrace which help us to know what internally is happening in the Solaris10 machine as I doubt my system configuration for the above

The way some DBs work is by having a separate lock monitor thread that runs every few seconds. This thread inspects a list of waiting locks for any cycles, which would indicate a circular relationship between processes holding locks & processes waiting for locks. The db then chooses a victim & kills one of the processes. (See "Inside MS SQL Server" by Delaney, pp778 for

I''ve found what I think is a deadlock in the dtrace fork code. My only Solaris test machine is a single cpu box, where I have been unable to reproduce in the field so far. I''ve been able to reproduce on other platforms. Here is the scenario. You''ve got a target process, which has static probes. It is calling fork(), which causes it to end up at line 330 in fork.c:

To remember all dtrace actions, aggregations and variables is a headache, even worse, they are not include in dtrace(1M) now. So here is a chance to have a CLI RFE - to have something like "dtrace --actions", which will list all actions, like this page: http://developers.sun.com/solaris/articles/dtrace_quickref/dtrace_functions.html "dtrace --aggregations", lists all

For those who may have missed the news, the DTrace movers and shakers have moved over to a new mailing list; you can sign up at http://dtrace.org/blogs/mailing-list/ And there will be a conference dedicated to DTrace on April 3rd in San Francisco, more information here: http://wiki.smartos.org/display/DOC/dtrace.conf -- best regards, Deirdr? Straughan SmartOS Community Manager Joyent cell

> > From: vikram b > > Date: October 22, 2009 12:31:53 AM PDT > > To: dtrace-discuss at opensolaris.org > > Subject: [dtrace-discuss] dtrace problem > > > > Hi All, > > I am running dtrace on linux !! got it working some how. > > dtrace -l shows a big list.. but many issues > > > > 1. says : failed to grab process > >

(Apologies to anyone who follows me on Twitter and therefore will have seen this multiple times already.) Joyent will be offering training in DTrace: Core Features and Performance Analysis trainer: Brendan Gregg <http://dtrace.org/blogs/brendan/> (San Francisco, Sept 6-7), Max Bruning <http://wiki.smartos.org/display/DOC/Max+Bruning> (Geneva, Sept 13-14) This course introduces

X.Org has adopted GNU autoconf as its build configuration mechanism, so when I integrated the dtrace probes, I checked to see if they should be built using this test, checking for the existence of a program named "dtrace" in the path: dnl Check for dtrace program (needed to build Xserver dtrace probes) AC_ARG_WITH(dtrace, AS_HELP_STRING([--with-dtrace=PATH], [Enable dtrace

Dear dtrace Experts, I have seen some dtrace utilities like opensnoop and execsnoop etc. My interest is to write a simple script that can snoop the files which uses the 3 syscalls like open,create,unlink. I have gone through dtrace oneliners that can do the same : dtrace -n ''syscall::open*:entry { printf("%s %s",execname,copyinstr(arg0));}'' dtrace -n

Have you considered the mdb dcmd ::pfiles? mfe at inker:~/Code/dtrace/examples$ *pfexec mdb -k* Loading modules: [ unix genunix specfs dtrace mac cpu.generic uppc pcplusmp scsi_vhci zfs ip hook neti sockfs arp usba uhci fctl s1394 stmf stmf_sbd md lofs random idm sd crypto fcp cpc fcip smbsrv nfs ufs logindmux ptm nsmb sppp nsctl sdbc rdc sv ii ipc ] > *::ps ! grep clock-applet* R 1744

The earlier post with the Dtrace probes definition (sent on April 23) was only to solicit feedback from all prospective users of iSCSI DTrace probes. I will follow up shortly with an updated document requesting approval from the community. thanks Priya -------- Original Message -------- Subject: Dtrace provider for the COMSTAR iscsi Target Port Provider - Review Request Date: Thu, 23 Apr

Hi, If you have a custom kernel (and therefore have duplicates of everything in /kernel in your custom kernel) and have noticed that when you try to use anonymous tracing, dtrace adds multiple copies of the forceload directives to /etc/system, e.g.: * vvvv Added by DTrace * * The following forceload directives were added by dtrace(1M) to allow for * tracing during boot. If these

The open solaris dtrace page (http://www.opensolaris.org/os/community/dtrace/) has a bad link to the DTrace Guide(http://docs.sun.com/db/doc/817-6223). This message posted from opensolaris.org

Hello, I''m trying to debug an issue with rpc.metad on Solaris 10, basecally I believe the failure path is - svc_init -- check_sys --- check_host ---- strcmp(hostname, mynode()) It seems that for some reason hostname and mynode didn''t match then I would like to use dtrace to show the argument of the strcmp However the dtrace pid provider can''t trace functions

everyone Here: How do you do! dtrace is interesting. I am studing how to write dtrace script and analyze situation. I read dtrace_example.pdf and 819-6959.pdf document more than five times with careful. because I am newbie to studing dtrace.so I still don''t know how to begin. for question: 1. What consumes time? 2. Which system call(s) consume(s) time?

Hello Srikant - A quantization distributes the results of your aggregation into ranges ordered by a power-of-two. Presumably what you''d do in your script is capture the inclusive elapsed time of each function call in your library, then use this quantization to see how tightly-banded the times are. Perhaps there''s some blocking I/O in some of your calls, for example, in which

Hi All, One of the customer is seeing the following error messages. He is using a S10U1 system with all latest patches. Dtrace script is attached. I am not seeing these error messages on S10U2 system. What could be the problem? Thanks, Gangadhar. ------------------------------------------------------------------------ rroberto at njengsunu60-2:~$ dtrace -AFs /lib/svc/share/kcfd.d dtrace:

Hi Nils, all, I tried the ''FBT-trick'' suggested below by Nils and (of course :-) ) it works and gives me a pile of output. Good enough to show that indeed Linux application behavior is not totally opaque, good enough to for instance show that ''system call mapping'' is taking place. I wonder, is there any documentation available for lxsyscall module ? To learn

Dtrace does work on some workstation, such as as follows, ----------------------------------------------------------- bash-3.00# dtrace -s ./mem.d `pgrep test` dtrace: script ''./mem.d'' matched 5 probes ^C bash-3.00# uname -a SunOS ftp 5.10 Generic_118833-17 sun4u sparc SUNW,Sun-Blade-100 bash-3.00# dtrace -V dtrace: Sun D 1.1

Hi, I am using anon dtrace followed with normal dtrace to collect the data of all fork,exec,exit . I found that dtrace could not be able to collect all the data. Lot of pids are missing. This happence when we boot the system and expect dtrace to collect all the fork , exec & exit it collects data but not all the pids are collected. Any inputs in this regard will be of great help. thanks