Difference between revisions of "Running Timer Code"
(update instructions -- toby & alex) |
|||
(12 intermediate revisions by 3 users not shown) | |||
Line 1: | Line 1: | ||
+ | ==Obtaining Kernel Timing Data== | ||
+ | Testing is now radically simplified with the <code>run_test.sh</code> script. | ||
+ | |||
+ | ===Setup a client and server with the tp_timer kernel=== | ||
+ | |||
+ | '''1. Build and install the kernel.''' | ||
+ | To compile the kernel, check out <code>llk-sources</code> from CVS and use a .config file from llk-sources/ccg-configs. Make sure that <code>CONFIG_TP_TIMER=y</code> (for tp_timer built in) or <code>CONFIG_TP_TIMER=m</code> (for a module version of tp_timer). Note that there is a significant chance that this compilation will crash the build host is running tp_timer (having the module unloaded is okay). This is probably a problem with NFS, so building it locally may work (but watch out for distcc and ccache that like to touch files in your home directory). In short, build on a machine that isn't running a tp_timer kernel. | ||
+ | |||
+ | You probably want the module config (<code><machine>-mod</code> below). The resulting kernel version will be '2.6.9-llk-mod.' After building, <code>./ikl.py</code> in llk-sources will install the kernel locally for you. 2.6.9-llk-mod is the default kernel for both grub (on bazaar) and yaboot | ||
+ | (on cairo). | ||
+ | |||
+ | <pre> | ||
+ | cvs co llk-sources | ||
+ | cvs co llk | ||
+ | cd llk-sources | ||
+ | cp ccg-configs/<machine>-mod .config # where machine is 'cairo' or 'bazaar' | ||
+ | make oldconfig | ||
+ | make -jN CC="ccache gcc" # where N is the number of hosts in DISTCC_HOSTS * 4 | ||
+ | ./ikl.py # ikl = install kernel locally | ||
+ | </pre> | ||
+ | |||
+ | '''2. Copy to another machine.''' | ||
+ | If you need to copy the kernel to another machine, run: | ||
+ | <pre> | ||
+ | llk/scripts/distribute.sh 2.6.9-llk-mod <host name> | ||
+ | </pre> | ||
+ | |||
+ | '''3. Manage your kernels.''' | ||
+ | We will probably accumulate a large number of kernels over the next couple days. These scripts will aid in managing that proliferation: | ||
+ | *<code>llk/scripts/kernel_store.sh</code> saves the currently running kernel binary + associated files to a local directory (default: /usr/local/kernels/<kernel name specified>). | ||
+ | *<code>llk/scripts/kernel_load.sh</code> restores one of the previously saved kernels to /boot (under the name of the currently running kernel). | ||
+ | |||
+ | '''Boot-loader info,''' incase it's needed (you shouldn't have to change this!): | ||
+ | <pre> | ||
+ | # on cairo: | ||
+ | vim /etc/yaboot.conf | ||
+ | ybin -v | ||
+ | |||
+ | # on bazaar: | ||
+ | vim /boot/grub/menu.lst | ||
+ | grub < /etc/grub.conf | ||
+ | </pre> | ||
+ | ===Check out and run the test script=== | ||
+ | This must be done on a non-testing node, e.g., hopper. If you run this on a node you're using to test and accidently touch NFS, the machine will crash. | ||
+ | <pre> | ||
+ | cvs co llk | ||
+ | cd llk/scripts | ||
+ | ./run_test.sh <client> <server> <protocol> <table name> <test set> | ||
+ | </pre> | ||
+ | Where <code><client></code> and <code><server></code> are short host names like '<code>b10</code>' and the protocol is '<code>udp</code>' or '<code>tcp</code>'. The script will load the module on both hosts, calibrate the test, run the test, fetch the data, unload the module, and insert the data into the database with the specified table name and test set label. | ||
+ | |||
+ | ==Older info== | ||
Basic Flow: | Basic Flow: | ||
#make sure the tp_timer module is installed and working on the client and server nodes | #make sure the tp_timer module is installed and working on the client and server nodes | ||
− | #clear /proc/tp_timer on both client and server machines (just read its contents; cat /proc/tp_timer) | + | #clear /proc/tp_timer on both client and server machines (just read its contents; cat /proc/tp_timer > /dev/null) |
#run pacgen on client and server to amass new data. | #run pacgen on client and server to amass new data. | ||
#read data into log files on both server and client | #read data into log files on both server and client | ||
#load data from the log files into database | #load data from the log files into database | ||
#look at data via php scripts | #look at data via php scripts | ||
− | |||
Machines setup with llk timing: b10,b11,c10,c11 | Machines setup with llk timing: b10,b11,c10,c11 | ||
Line 12: | Line 63: | ||
Need to check out from CVS: | Need to check out from CVS: | ||
* llk-pacgen - cvs module with modified pacgen | * llk-pacgen - cvs module with modified pacgen | ||
− | |||
− | |||
− | |||
− | |||
− | |||
pacgen sample commands: | pacgen sample commands: |
Latest revision as of 05:43, 21 February 2006
Contents
Obtaining Kernel Timing Data
Testing is now radically simplified with the run_test.sh
script.
Setup a client and server with the tp_timer kernel
1. Build and install the kernel.
To compile the kernel, check out llk-sources
from CVS and use a .config file from llk-sources/ccg-configs. Make sure that CONFIG_TP_TIMER=y
(for tp_timer built in) or CONFIG_TP_TIMER=m
(for a module version of tp_timer). Note that there is a significant chance that this compilation will crash the build host is running tp_timer (having the module unloaded is okay). This is probably a problem with NFS, so building it locally may work (but watch out for distcc and ccache that like to touch files in your home directory). In short, build on a machine that isn't running a tp_timer kernel.
You probably want the module config (<machine>-mod
below). The resulting kernel version will be '2.6.9-llk-mod.' After building, ./ikl.py
in llk-sources will install the kernel locally for you. 2.6.9-llk-mod is the default kernel for both grub (on bazaar) and yaboot
(on cairo).
cvs co llk-sources cvs co llk cd llk-sources cp ccg-configs/<machine>-mod .config # where machine is 'cairo' or 'bazaar' make oldconfig make -jN CC="ccache gcc" # where N is the number of hosts in DISTCC_HOSTS * 4 ./ikl.py # ikl = install kernel locally
2. Copy to another machine. If you need to copy the kernel to another machine, run:
llk/scripts/distribute.sh 2.6.9-llk-mod <host name>
3. Manage your kernels. We will probably accumulate a large number of kernels over the next couple days. These scripts will aid in managing that proliferation:
llk/scripts/kernel_store.sh
saves the currently running kernel binary + associated files to a local directory (default: /usr/local/kernels/<kernel name specified>).llk/scripts/kernel_load.sh
restores one of the previously saved kernels to /boot (under the name of the currently running kernel).
Boot-loader info, incase it's needed (you shouldn't have to change this!):
# on cairo: vim /etc/yaboot.conf ybin -v # on bazaar: vim /boot/grub/menu.lst grub < /etc/grub.conf
Check out and run the test script
This must be done on a non-testing node, e.g., hopper. If you run this on a node you're using to test and accidently touch NFS, the machine will crash.
cvs co llk cd llk/scripts ./run_test.sh <client> <server> <protocol> <table name> <test set>
Where <client>
and <server>
are short host names like 'b10
' and the protocol is 'udp
' or 'tcp
'. The script will load the module on both hosts, calibrate the test, run the test, fetch the data, unload the module, and insert the data into the database with the specified table name and test set label.
Older info
Basic Flow:
- make sure the tp_timer module is installed and working on the client and server nodes
- clear /proc/tp_timer on both client and server machines (just read its contents; cat /proc/tp_timer > /dev/null)
- run pacgen on client and server to amass new data.
- read data into log files on both server and client
- load data from the log files into database
- look at data via php scripts
Machines setup with llk timing: b10,b11,c10,c11
Need to check out from CVS:
- llk-pacgen - cvs module with modified pacgen
pacgen sample commands:
./server tcp <port number> ./client -t <protocol> -s <source IP> -m <source port> -d <dest IP> -p <dest port>
Each window is a specific ID for the packet. This ID is used as the payload for the packets. Useful for checking for packet fragmentation (educated guess via Alex). Windows are 0xffffffff, 0xffffffff (2 unsigned ints worth of 1's), an unsigned int for socket/thread id (We don't use it; that part of the code was removed), and an unsigned int holding the window number (we don't know what this is for).
After pacgen is ran with on a tp timer enabled machine, the timing results will be available in /proc/tp_timer. Note that after you read this proc entry, the information will be cleared.
To enter the data into our database:
- Collect the /proc/tp_timer from the server node and the client node:
- cat /proc/tp_timer > tp_timer_client.log
- Execute the create-table-llk.pl script found in the the llk cvs module in the scripts sub directory.
- Usage: ./create-table-llk.pl <tablename> <label> <datafile> <client|server>
- If the table given by name to the scripts doesn't exist, the script creates the table.