I manage a large number of shell (ksh) scripts on server A. Each script begins with the line...
#!/usr/bin/ksh
When I deploy to machine B, C, and D I frequently need to use a different shell such as /bin/ksh, /usr/local/bin/ksh or even /usr/dt/bin/ksh. Assume I am unable to install a new version of ksh and I am unable to create links in any protected directories such as /usr/local/bin. At the moment I have a sed script which modifies all the scripts but I would prefer not to do this. I would like to standardize the header so that it no longer needs to be changed from server to server. I don't mind using something like
#!~/ksh
And creating a link which is on every server but I have had problems with finding home using "~" in the past when using rsh (maybe is was ssh) to call a script (AIX specifically I think). Another option might be to create a link in my home directory and ensuring that it is first in my PATH, and simply using
#!ksh
Looking for a good solution. Thanks.
Update 8/26/11 - Here is the solution I came up with. The installation script looks for the various versions of ksh installed on the server and then copies one of the ksh 93 programs to /tmp/ksh93. The scripts in the framework all refer to #!/tmp/ksh93 and they don't need to be changed from one server to the other. The script also set some variables so that if the file is every removed from /tmp, it will immediately be put back the next time a scheduled task runs, which is at a minimum every minute.
As rettops noted, you can use:
#!/usr/bin/env ksh
This will likely work for you. However, there can be some drawbacks. See Wikipedia on Shebang for a fairly thorough discussion.
#! /usr/bin/env ksh
will use whatever ksh is in the user's path.
Related
I'm working on cluster and using custom toolkits (more specifically SRA toolkit). In order to use it, I fist had to download (and unpack it) to a specific folder in my directory.
Then I had to modify .bashsrc to include the following segment:
# User specific aliases and functions
export PATH="$PATH:/home/MYNAME/APPS/SRATOOLS/bin"
Now I can use a stuff from SRATools in bash command line, e.g.
prefetch SR111111
My question is, can I use those tools without modifying my .bashsrc?
The reason that I want to do that is because I wrote a .sh script that takes a long time to run, and my cluster has Sun Grid Engine job management system, and I submitted my script to it, only to see the process fail - because a SRA Toolkit command I used was unrecognized.
EDIT (1):
I modified the location where my prefetch command is, and now it looks like:
/MYNAME/APPS/SRA_TOOLS/bin
different from how it is in .bashsrc:
export PATH="$PATH:/home/MYNAME/APPS/SRATOOLS/bin"
And run what #Darkman suggested (put IF THEN ELSE FI and under ELSE put export). The output is that it didn't find SRATools (because path in .bashsrc is different), but it found them under ELSE and script is running normally. Weird. It works on my job management system.
Thanks everybody.
I ask because I recently made a change to a KornShell (ksh) script that was executing. A short while after I saved my changes, the executing process failed. Judging from the error message, it looked as though the running process had seen some -- but not all -- of my changes. This strongly suggests that when a shell script is invoked, the entire script is not read into memory.
If this conclusion is correct, it suggests that one should avoid making changes to scripts that are running.
$ uname -a
SunOS blahblah 5.9 Generic_122300-61 sun4u sparc SUNW,Sun-Fire-15000
No. Shell scripts are read either line-by-line, or command-by-command followed by ;s, with the exception of blocks such as if ... fi blocks which are interpreted as a chunk:
A shell script is a text file containing shell commands. When such a
file is used as the first non-option argument when invoking Bash, and
neither the -c nor -s option is supplied (see Invoking Bash), Bash
reads and executes commands from the file, then exits. This mode of
operation creates a non-interactive shell.
You can demonstrate that the shell waits for the fi of an if block to execute commands by typing them manually on the command line.
http://www.gnu.org/software/bash/manual/bashref.html#Executing-Commands
http://www.gnu.org/software/bash/manual/bashref.html#Shell-Scripts
It's funny that most OS'es I know, do NOT read the entire content of any script in memory, and run it from disk. Doing otherwise would allow making changes to the script, while running. I don't understand why that is done, given the fact :
scripts are usually very small (and don't take many memory anyway)
at some point, and shown in this thread, people would start making changes to a script that is already running anyway
But, acknowledging this, here's something to think about: If you decided that a script is not running OK (because you are writing/changing/debugging), do you care on the rest of the running of that script ? you can go ahead making the changes, save them, and ignore all output and actions, done by the current run.
But .. Sometimes, and that depends on the script in question, a subsequent run of the same script (modified or not), can become a problem since the current/previous run is doing an abnormal run. It would typically skip some stuff, or sudenly jump to parts in the script, it shouldn't. And THAT may be a problem. It may leave "things" in a bad state; particularly if file manipulation/creation is involved.
So, as a general rule : even if the OS supports the feature or not, it's best to let the current run finish, and THEN save the updated script. You can change it already, but don't save it.
It's not like in the old days of DOS, where you actually have only one screen in front of you (one DOS screen), so you can't say you need to wait on run completion, before you can open a file again.
No they are not and there are many good reasons for that.
One of the things you should keep in mind is that a shell is not an interpreter even if there are some similarities. Shells are designed to work with a stream of commands. Either from the TTY ,a PIPE, FIFO or even a socket.
The shell reads from its resource line by line until a EOF is returned by the kernel.
The most shells have no extra support for interpreting files. they work with a file as they would work with a terminal.
In fact this is considered to be a nice feature because you can do interesting stuff like this How do Linux binary installers (.bin, .sh) work?
You can use a binary file and prepend shell scripts. You can't do this with an interpreter. because it parses the whole file or at least it would try it and fail. A shell would just interpret it line by line and doesnt care about the garbage at the end of the file. You just have to make sure the execution of the script gets terminated before it reaches the binary part.
I have a node.js app which is using the child_process.execFile command to run a command-line utility.
I'm worried that it would be possible for a user to run commands locally (a rm / -rf horror scenario comes to mind).
How secure is using execFile for Bash scripts? Any tips to ensure that flags I pass to execFile are escaped by the unix box hosting the server?
Edit
To be more precise, I'm more wondering if the arguments being sent to the file could be interpreted as a command and executed.
The other concern is inside the bash script itself, which is technically outside the scope of this question.
Using child_process.execFile by itself is perfectly safe as long as the user doesn't get to specify the command name.
It does not run the command in a shell (like child_process.exec does), so there is no need to escape anything.
child_process.execFile will execute commands with the user id of the node process, so it can do anything that user could do, which includes removing all the server files.
Not a good idea to let user pass in command as you seem to be implying by your question.
You could consider running the script in a sandbox by using chroot, and limiting the commands and what resides on the available file system, but this could get complet in a hurry.
The command you pass will get executed directly via some flavor of exec, so unless what you trying to execute is a script, it does not need to be escaped in any way.
While playing with Heroku, I found their approach of using environment variables for server-local configuration brilliant. Now, while setting up an application server of my own, I find myself wondering how hard that would be to replicate.
I'm deploying a sinatra application, riding Unicorn and Nginx. I know nginx doesn't like to play with the environment, so that one's out. I can probably put the vars somewhere in the unicorn config file, but since that's under version control with the rest of the app, it sort of defeats the purpose of having the configuration sit in the server environment. There is no reason not to keep my app-specific configuration files together with the rest of the app, as far as I'm concerned.
The third, and last (to my knowledge) option, is setting them in the spawning shell. That's where I got lost. I know that login and non-login shells use different rc files, and I'm not sure whether calling something with sudo -u http stuff is or not spawning a login shell. I did some homework, and asked google and man, but I'm still not entirely sure on how to approach it. Maybe I'm just being dumb... either way, I'd really appreciate it if someone could shed some light on the whole shell environment deal.
I think your third possibility is on the right track. What you're missing is the idea of a wrapper script, whose only function is to set the environment and then call the main program with whatever options are required.
To make a wrapper script that can function as a control script (if prodEnv use DB=ProdDB, etc), there is one more piece that simplifies this problem. Bash/ksh both support a feature called sourcing files. This an operation that the shell provides, to open a file and execute what is in the file, just as if it was in-lined in the main script. Like #include in C and other languages.
ksh and bash will automatically source /etc/profile, /var/etc/profile.local (sometimes), $HOME/.profile. There are other filenames that will also get picked up, but in this case, you'll need to make your own env file and the explicitly load it.
As we're talking about wrapper-scripts, and you want to manage how your environment gets set up, you'll want to do the sourcing inside the wrapper script.
How do you source an environment file?
envFile=/path/to/my/envFile
. $envFile
where envFile will be filled with statements like
dbServer=DevDBServer
webServer=QAWebServer
....
you may discover that you need to export these variable for them to be visble
export dbServer webServer
An alternate assignment/export is supported
export dbServer=DevDBServer
export webServer=QAWebServer
Depending on how non-identical your different environments are, you can have your wrapper script figure out which environment file to load.
case $( /bin/hostame ) in
prodServerName )
envFile=/path/2/prod/envFile ;;
QASeverName )
envFile=/path/2/qa/envFile ;;
devSeverName )
envFile=/path/2/dev/envFile ;;
esac
. ${envFile}
#NOW call your program
myProgram -v -f inFile -o outFile ......
As you develop more and more scripts in your data processing environment, you can alway source your envFile at the top. When you eventually change the physical location of a server (or it's name), then you have only one place that you need to make the change.
IHTH
Also a couple of gems dealing with this. figaro that works both with or without heroku. Figaro uses a yaml file (in config and git ignored) to keep track of variables. Another option is dotenv that reads variables from an .env file. And also another article with all them options.
To spawn an interactive shell (a.k.a. login shell) you need to invoke sudo like this:
sudo -i -u <user> <command>
Also you may use -E to preserve the environment. This will allow some variables to be pased for your current environment to the command invoked with sudo.
I solved a similar problem by explicitly telling Unicorn to read a variables file as part of startup in its init.d script. First I created a file in a directory above the application root called variables. In this script I call export on all my environment variables, e.g. export VAR=value. Then I defined a variable GET_VARS=source /path/to/variables in the /etc/init.d/unicorn file. Finally, I modified the start option to read su - $USER -c "$GET_VARS && $CMD" where $CMD is the startup command and $USER is the app user. Thus, the variables defined in the file are exported into the shell of Unicorn's app user on startup. Note that I used an init.d script almost identical to the one from this article.
I have a ruby script containing system command like http://gist.github.com/235833, while I ran this script from shell, it works correctly, but when I added it to my cron job list, it doesn't work any more, the cron job is like:
10/* * * * * cd /home/hekin; /usr/bin/ruby my_script.rb
any idea what's going wrong with what i've done? Thank you.
Thank you all for your answers.
It's my mistake.
Since I'm using ssh key forwarding on the local machine, while I executed the script from the shell, the ssh key forwarding related environment variables are all sitting there, but from cron job context, those environment variables are missing.
Try to separate the things that might go wrong. The ones I can think of are:
The cron syntax - is the time value given legal and fitting your shell?
Permissions - execute permissions and read permissions for the relevant directory and file
Quoting - what scope does cron cover? Does it run only the first command?
In order to dissect this, I suggest you first run a really simple cron job, like 'ls'. Next run a single-liner script. Next embed your commands in a shell-script file. Somewhere along these lines you should find the problem.
The problem is your environment. While testing in your shell its fully equipped and boosted by your shell environment. While running under cron its very, very stripped down.
Where is the destination "." for your script? I guess it will be "/" and may not "$HOME" thus your script won't be able to write at that location and fails. Try using an absolut path for the destination.