At the CommunityOne West keynote last June a demo of Crossbow was given using a tool called Virtual Wire. If you haven't seen the demo before, you can watch a 10 minute replay of it here:

Since that time a lot of folks have asked for access to the tool that was used in the demo. Browsing the Crossbow project page recently I noticed the Crossbow Virtual Wire Demo Tool has now been made available.

I'll warn you in advance that I view the tool as alpha quality as it doesn't have some basic functionalities like undo. However, as the names implies, it's perfect for demos (and learning a lot about zones and networking).

When you extract the tool you'll find a couple of screencasts, which I've also made available here. There's no audio, but they do give you the basic idea of the capabilities of the tool.

Demo 1: Creating and Connecting 2 Private Networks

This demo uses the Crossbow Virtual Wire Demo Tool to create 2 private networks. One network contains a web server (192.168.2) while the client is connecting from a separate network (193.168.3). A router is set up to forward packets between the 2 networks. This is a self-contained private network. The demo concludes by using a network load generator along with DTrace to show how easy it is to monitor and configure bandwidth to the server. The details on how to reproduce this demo have been documented here.

Demo 2: Creating a Private Network Connected to the Public Network

This demo uses the Crossbow Virtual Wire Demo Tool to create a private network containing a web server behind a firewall. The IP routing is configured to access the server from the global zone. The demo concludes by using a network load generator along with DTrace to show how easy it is to monitor and configure bandwidth to the server. The details on how to reproduce this demo have been documented here.

Read more >>

I've read that the best way to learn DTrace is to study existing D scripts. However, upon doing so I quickly find myself asking "where is that documented?" And it turns out, with tens of thousands of probes, as is often the case, it's not. So I'd love to hear what others with more DTrace experience do, but these are the steps I seem to have stumbled upon.

As an example, let's take the script /opt/DTT/Proc/rwbytype.d, which shows the bytes read and written by process by vnode type. Here's an example of the output:

bleonard@opensolaris:/opt/DTT/Proc$ rwbytype.d 
Tracing... Hit Ctrl-C to end.
^C
PID    CMD               VTYPE  DIR     BYTES
21124  gnome-terminal      chr    W         1
10523  gnome-power-mana   sock    W         4
10565  gnome-netstatus-   fifo    R         7
10547  gnome-netstatus-   fifo    W         8
4817   conky              fifo    R         9
10634  firefox-bin        fifo    R         9
10634  firefox-bin        fifo    W         9
15379  whoami             fifo    W         9
10547  gnome-netstatus-   fifo    R        20
10565  gnome-netstatus-   fifo    W        20
15377  dtrace              chr    W        31
10412  Xorg                chr    R        32
10469  gnome-settings-d   sock    R        32
10523  gnome-power-mana   sock    R        32
21124  gnome-terminal      chr    R        33
10472  metacity           sock    W        56
10472  metacity           sock    R       200
21124  gnome-terminal     sock    R       220
15379  conky               reg    R       529
15379  ksh93               reg    R       529
4817   conky              sock    R      2720
10412  Xorg               sock    W      3204
4817   conky               reg    R      6696
21124  gnome-terminal     sock    W      6852
4817   conky              sock    W     37952
10412  Xorg               sock    R     44864

And here's the D script that produces that output:

typedef struct vtype2str {
	string code;
};

translator struct vtype2str < int T > {
	/* the order has been picked for performance reasons */
	code =
	    T == 1 ? "reg"   :
	    T == 9 ? "sock"  :
	    T == 4 ? "chr"   :
	    T == 6 ? "fifo"  :
	    T == 8 ? "proc"  :
	    T == 2 ? "dir"   :
	    T == 3 ? "blk"   :
	    T == 5 ? "lnk"   :
	    T == 7 ? "door"  :
	    T == 10 ? "port" :
	    T == 11 ? "bad"  : "non";
};

dtrace:::BEGIN
{
	printf("Tracing... Hit Ctrl-C to end.\n");
}

fbt::fop_read:entry,
fbt::fop_write:entry
{
	self->type = xlate <struct vtype2str *>(args[0]->v_type)->code;
	self->size = args[1]->uio_resid;
	self->uiop = args[1];
}

fbt::fop_read:return
/self->uiop/
{
	this->resid = self->uiop->uio_resid;
	@bytes[pid, execname, self->type, "R"] = sum(self->size - this->resid);
	self->type = 0;
	self->size = 0;
	self->uiop = 0;
}

/* this is delibrately redundant code for performance reasons */
fbt::fop_write:return
/self->uiop/
{
	this->resid = self->uiop->uio_resid;
	@bytes[pid, execname, self->type, "W"] = sum(self->size - this->resid);
	self->type = 0;
	self->size = 0;
	self->uiop = 0;
}

dtrace:::END
{
	printf("%-6s %-16s %6s %4s %9s\n",
	    "PID", "CMD", "VTYPE", "DIR", "BYTES");
	printa("%-6d %-16s %6s %4s %@9d\n", @bytes);
}

First of all, this D script is only enabling 6 probes:

dtrace:::BEGIN
dtrace:::END
dtrace::fop_read:entry
dtrace::fob_read:return
dtrace::fob_write:entry
dtrace::fob_write:return

The BEGIN and END clauses are self-explanatory. Where I get tripped up is looking at the 2nd clause:

fbt::fop_read:entry,
fbt::fop_write:entry
{
	self->type = xlate <struct vtype2str *>(args[0]->v_type)->code;
	self->size = args[1]->uio_resid;
	self->uiop = args[1];
}

First, I see that we're setting 3 thread-local variables, type, size and uiop. All of these variables are set using probe arguments, and this is where things start to get confusing - where are these arguments defined? I can run the following dtrace command to get verbose information on a probe:

bleonard@opensolaris:/opt/DTT/Proc$ dtrace -lv -n fbt::fop_read:entry
   ID   PROVIDER            MODULE                          FUNCTION NAME
17725        fbt           genunix                          fop_read entry

	Probe Description Attributes
		Identifier Names: Private
		Data Semantics:   Private
		Dependency Class: Unknown

	Argument Attributes
		Identifier Names: Private
		Data Semantics:   Private
		Dependency Class: ISA

	Argument Types
		args[0]: vnode_t *
		args[1]: uio_t *
		args[2]: int
		args[3]: cred_t *
		args[4]: caller_context_t *

So I learn that the data semantics of the argument attributes are private, which means this probe is accessing private implementation details of the kernel, but that's the case with most DTrace kernel probes. Of more interest is the argument list:

args[0]: vnode_t *
args[1]: uio_t *
args[2]: int
args[3]: cred_t *
args[4]: caller_context_t 

So I can see there are 5 total arguments available for this probe and I know my clause is accessing args[0] and args[1]. But what now? What exactly is vnode_t and uio_t? As far as I know, there's no further dtrace command to drill down into these data types. At this point I turn to the OpenSolaris Source Browser.

I begin by searching on the probe function fop_read, which returns 2 results, the vnode.h header file and the vnode.c source file:

Searched refs:fop_read (Results 1 -2 of 2) sorted by null

/onnv/onnv-gate/usr/src/uts/common/sys/
vnode.h	[all...]
/onnv/onnv-gate/usr/src/uts/common/fs/
vnode.c	3177 fop_read( function     [all...]

The fop_read function is defined on line 3177 of vnode.c and I can click on that link to be taken directly to the source. Sure enough, the arguments to the fop_read function in vnode.c match what I was told by dtrace:

   3176 int
fop_read(
	vnode_t *vp,
	uio_t *uiop,
	int ioflag,
	cred_t *cr,
	caller_context_t *ct)

Now I can click on vnode_t to see its definition and this returns 3 results - the logical one being the vnode.h header file returned by our first search:

/onnv/onnv-gate/usr/src/uts/common/sys/ vnode.h 257 } vnode_t; typedef in typeref:struct:vnode 800 int (*vop_open)(vnode_t **, int, cred_t *, \ 802 int (*vop_close)(vnode_t *, int, int, offset_t, cred_t *, \ 804 int (*vop_read)(vnode_t *, uio_t *, int, cred_t *, \ 806 int (*vop_write)(vnode_t *, uio_t *, int, cred_t *, \ 808 int (*vop_ioctl)(vnode_t *, int, intptr_t, int, cred_t *, \ 810 int (*vop_setfl)(vnode_t *, int, int, cred_t *, \ 812 int (*vop_getattr)(vnode_t *, vattr_t *, int, cred_t *, \ 814 int (*vop_setattr)(vnode_t *, vattr_t *, int, cred_t *, \ 816 int (*vop_access)(vnode_t *, int, int, cred_t *,     [all...]

Here we see that vnode_t is defined on line 257, which you can click to navigate directly to that section of the source. Our D script above is referencing args[0]->v_type, and we can see v_type is an enum:

typedef struct vnode {
	kmutex_t	v_lock;		/* protects vnode fields */
	uint_t		v_flag;		/* vnode flags (see below) */
	uint_t		v_count;	/* reference count */
	void		*v_data;	/* private data for fs */
	struct vfs	*v_vfsp;	/* ptr to containing VFS */
	struct stdata	*v_stream;	/* associated stream */
	enum vtype	v_type;		/* vnode type */
	dev_t		v_rdev;		/* device (VCHR, VBLK) */

Which you can also click on to see the type definition:

/*
 * vnode types.  VNON means no type.  These values are unrelated to
 * values in on-disk inodes.
 */
typedef enum vtype {
	VNON	= 0,
	VREG	= 1,
	VDIR	= 2,
	VBLK	= 3,
	VCHR	= 4,
	VLNK	= 5,
	VFIFO	= 6,
	VDOOR	= 7,
	VPROC	= 8,
	VSOCK	= 9,
	VPORT	= 10,
	VBAD	= 11
} vtype_t;

Which matches up perfectly to the struct at the beginning of our D script. So, now I've got a better handle of exactly what's going on with this D script.

Now, as a test, see if you can follow the same procedure to find the documented comment in the OpenSolaris source code for arg[1]->uio_resid :-). 

Read more >>

Sun ISV engineer and DTrace evangelist Angelo Rajadurai has just authored a new quick start guide to DTrace. The guide is based on his "Observing Applications with DTrace" presentation (slides, video) at CommunityOne West and focuses on tracing web applications. The examples from the book have also been added to a wiki page for easier copy & paste.

Read more >>

A hidden little gem that comes included with your installation of OpenSolaris is the DTrace Toolkit. You'll find it installed at /opt/DTT.

The DTrace Toolkit is a collection of tools written by volunteers of the DTrace community, led by Brendan Gregg. What I find most impressive is how well it has been put together. There are 230 executable DTrace scripts, each with their own man page. The scripts are organized by type (CPU analysis, I/O analysis, etc.) so they are easy to study, but they've all been symlinked into a Bin directory, so they are easy to put on your path. I've added the following to my ~/.bashrc file:

#
# DTrace Toolkit
#
PATH=$PATH:/opt/DTT/Bin
MANPATH=$MANPATH:/opt/DTT/Man

In addition to their man pages, all of the scrips have a -h help option, for example:

bleonard@os200906:~$ execsnoop -h
USAGE: execsnoop [-a|-A|-ehjsvZ] [-c command]
       execsnoop                # default output
                -a              # print all data
                -A              # dump all data, space delimited
                -e              # safe output, parseable
                -j              # print project ID
                -s              # print start time, us
                -v              # print start time, string
                -Z              # print zonename
                -c command      # command name to snoop
  eg,
        execsnoop -v            # human readable timestamps
        execsnoop -Z		# print zonename
        execsnoop -c ls         # snoop ls commands only

Note, if you want to run the DTrace scripts from your user account, run the following command to add the needed privileges, substituting in your user name:

pfexec usermod -K defaultpriv=basic,dtrace_user,dtrace_proc,dtrace_kernel <user>

To get started learning all that's available, I recommend reading the DTrace Tookkit page. Also, Stefan Parvu has an excellent presentation on DTrace and the toolkit that's worth a look.

With this toolkit there's probably little need to write a DTrace script from scratch - it's literally a treasure chest of material that I've only begun to explore.

Read more >>

There are (at least) 2 big Linux conferences this weekend, the Ontario Linux Fest and the Florida Linux Show. The good news is that you'll find OpenSolaris represented at both of them. In Ontario Canada, Toronto OpenSolaris User Group leader Steven Acres will deliver Why OpenSolaris? In Orlando, Florida, I'll be delivering What's So Cool About OpenSoalaris Anyhow?

If you happen to be nearby (or willing to travel), please stop by and say hello.

Read more >>

When I edit a text file in vi, I find it much easier to read if the syntax is color highlighted. There are 2 steps to enabling this. First, add an alias for vi (or vim) to your ~/.bashrc file as follows:

#
# Enable color support for vi
#
alias vi='TERM=xtermc vim'

You can read the gory details about why this needs to be done in bug 1641 - xterm terminfo entry doesn't support color.

You also need to tell vi to use syntax coloring. Adding the following to your ~/.vimrc file (create one if it doesn't yet exist):

syntax on

Much nicer:

Read more >>

Netperf is a cool little utility that I discovered while working with Crossbow. It's technically a network benchmarking tool, but it's fun to use to hammer your network with load and test out Crossbow's bandwidth control features.

Simply click the link to install netperf , which will launch the Package Manager Web Installer:

To verify the installation, first start the benchmark server:

bleonard@opensolaris:~$ netserver
Starting netserver at port 12865
Starting netserver at hostname 0.0.0.0 port 12865 and family AF_UNSPEC

Then run netperf, which will run a TCP_STREAM test of 10 seconds over the loopback interface:

bleonard@opensolaris:~$ netperf
TCP STREAM TEST from ::ffff:0.0.0.0 (0.0.0.0) port 0 AF_INET to opensolaris (127.0.0.1) port 0 AF_INET
Recv   Send    Send                          
Socket Socket  Message  Elapsed              
Size   Size    Size     Time     Throughput  
bytes  bytes   bytes    secs.    10^6bits/sec  

 49152  49152  49152    10.00    4895.46  

Now that we know Netperf is set up correctly, we can use it to test bulk data transfer performance between 2 hosts (or 2 zones). TCP stream performance is the default test type and it's simply performed by running the command:

netperf -H <remotehost>

which will run a 10 second test between the local and remote system.

So I'm going to use netperf to run some tests between the global zone and a local zone, which has been set up to use a Crossbow virtual NIC exclusively. To create such a configuration on your own, see the blog entry Zones and Crossbow.

The local zone, myzone, is currently up and running:

bleonard@opensolaris:~$ zoneadm list -cv
  ID NAME             STATUS     PATH                           BRAND    IP    
   0 global           running    /                              native   shared
   - myzone           running    /zones/myzone                  ipkg     excl 

The first thing I'm going to do is copy the netserver program over to the local zone:

pfexec cp /opt/csw/bin/netserver /zones/myzone/root/var/tmp/.

And then start the netserver:

bleonard@opensolaris:~$ pfexec zlogin myzone /var/tmp/netserver
en_US.UTF-8: unknown locale
en_US.UTF-8: unknown locale
Starting netserver at port 12865
Starting netserver at hostname 0.0.0.0 port 12865 and family AF_UNSPEC

Now let's test the connection between the global and local zone:

bleonard@opensolaris:~$ netperf -H 10.0.1.25
TCP STREAM TEST from ::ffff:0.0.0.0 (0.0.0.0) port 0 AF_INET to ::ffff:10.0.1.25 (10.0.1.25) port 0 AF_INET
Recv   Send    Send                          
Socket Socket  Message  Elapsed              
Size   Size    Size     Time     Throughput  
bytes  bytes   bytes    secs.    10^6bits/sec  

 49152  49152  49152    10.00    1130.01   

Here we can see the throughput between my zones is 1130 Mbit/s. Now let's reduce the max bandwidth of the virtual NIC to 500 Mbit/s and try the test again:

bleonard@opensolaris:~$ pfexec dladm set-linkprop -p maxbw=500 myzone0
bleonard@opensolaris:~$ dladm show-vnic
LINK         OVER         SPEED  MACADDRESS           MACADDRTYPE         VID
myzone0      e1000g0      500    2:8:20:59:0:b5       random              0
bleonard@opensolaris:~$ netperf -H 10.0.1.25
TCP STREAM TEST from ::ffff:0.0.0.0 (0.0.0.0) port 0 AF_INET to ::ffff:10.0.1.25 (10.0.1.25) port 0 AF_INET
Recv   Send    Send                          
Socket Socket  Message  Elapsed              
Size   Size    Size     Time     Throughput  
bytes  bytes   bytes    secs.    10^6bits/sec  

 49152  49152  49152    10.00     486.22   

Or at a ridiculously low 2 Mbit/s:

bleonard@opensolaris:~$ pfexec dladm set-linkprop -p maxbw=2 myzone0
bleonard@opensolaris:~$ netperf -H 10.0.1.25
TCP STREAM TEST from ::ffff:0.0.0.0 (0.0.0.0) port 0 AF_INET to ::ffff:10.0.1.25 (10.0.1.25) port 0 AF_INET
Recv   Send    Send                          
Socket Socket  Message  Elapsed              
Size   Size    Size     Time     Throughput  
bytes  bytes   bytes    secs.    10^6bits/sec  

 49152  49152  49152    10.25       1.09   

Good stuff.

I've just scratched the surface of Netperf, but this simple introduction suits my purposes for testing Crossbow. For more fun check out the Netperf manual.

Read more >>

I've had my local IPS repository mirror running for about 1 month now and I find it extremely beneficial. So much so that I've gone ahead and produced a short 6 minute screencast on the process of creating it (although it takes much longer than that to set up as you need to wait for files to download and copy - the joys of editing screencasts).

 Click the image to play the screencast

Read more >>