Solaris 8 NIS Implementation

I'm attempting to determine if there is a problem with
SunOS 5.8 Generic_108528-11. The problem :
When using the /usr/sbin/ypinit script to create a master domain/ slave domain
I keep getting kicked out due to domainname not being defined. The domain is
defined and can be echo'ed during the ypinit execution. To resolve I comment
out the error routine that kicks me out of the script.
Also, /usr/lib/netsvc/yp/multi - had to set the domainname within the script.
Also, if anybody knows of any site that has some details/theory about NIS
please email me at [email protected]
Thanks,
Marv

I don't know about the second part but you might want to look at the bool "Java Threads 2nd edition" for the answer to the first part.

Similar Messages

Solaris nis client can't connect to Windows 2008 NIS Server across subnets

I have been using Microsoft Windows Server for NIS for years to centralize some accounts accessing Solaris and Linux OS.
Windows 2003 R2 Indentity Management for UNIX version of Server for NIS was last working version with Solairs.
After upgrading Domain to Windows 2008 all Solaris clients that are on different subnets fail to bind and connect to NIS servers now.
These servers already were setup using ypinit -c to manually add the NIS servers to connect to and prior to Windows 2008 this worked.
Now no Solaris nis clients on different subnets (no firewall between) can connect , but other Unix/Linux OS connect fine.
Is this a known issue and can we get a resolution as it seems only Solaris is affected?

Hi Mid.Hudson-IT,
Before we begin ,we should ensure we have configured the printer server correctly .
Here is a link for reference of configuring the printer server .
Print server role: Configuring a print server
https://technet.microsoft.com/en-us/library/cc775791(v=ws.10).aspx
"I can map all of the printers on the network to the print server as-well as install the drivers correctly and I can resolve the IP and server name via DNS"
From this sentence ,I can`t figure out whether you have tried to ping the server both with the IP adress and name adress from the client ?
If we can ping the print server from the client,we can ensure the connection to the printer server is good .
Then we can try to install the printer driver directly to have a check .In the adress bar of Windows Explorer ,input "\\server name \the printer name"
If we cannot ping the print server ,we should troubleshoot the network issue firstly.
We also can check the event viewer for more information to troubleshoot this issue .
Best regards

Where is Solaris OE Implementation for Win32 API Version 1.0.2.?

the link on page http://soldc.sun.com/ntmigration/tools/jscoretool.html is broken.
I search and found with no results.
can anyone help ? or send a copy to my email box: [email protected]
thanks a lot.

>
Try again for olaris[tm] OE Analyzer for C/C++ Source
Code at
http://soldc.sun.com/ntmigration/tools/jscoretool.html
It has a new version now as 1.1.
Also the Solaris OE Implementation for Win32 APIs
also has a new version at
http://soldc.sun.com/ntmigration/nttech/win32emulation
.html
Enjoy!The above link is broken. Can anyone please mail me the valid working link at [email protected]
Would really appreciate your help,
Thanks and Regards,
Sanjay

Managing NIS accounts on Solaris resource

Perhaps You have some helpful experience managing Solaris NIS accounts ?
I'm using IDM 8.0 and observe a quite disappointing situation:
Reconciliation of about 5000 accounts takes 10hours!
1) Resource Reference suggests to improve the performance of "logins" command
on Solaris by installing patch 126632-01, which does not exist:-/
The patch 125549-01 (found in sunsolve) requires Solaris 10, my scenario has Solaris 9:-/
Has anybody installed that patch and got better performance then ?? How many percent speed-up - 1,10,90 ?
2) To "increase performance during bulk provisioning" Resource Reference also suggests
to add and use (??) an attribute "user_make_nis" (same name on resource and IDM ?)
and/or (??) "create a ResourceAction named NIS_password_make in the workflow" !?
Are these 2 features meant to work each independent or only together ?
Did anybody try to do that - with noticeable effect ?
During "reconciliation" in general or only "user update on the resource" (provisioning) ?
What means to "use" the attribute and what is content of the ResourceAction (cd /var/yp; make)?
3) Do You have other suggestions for improvement (based on experience) ?
4) Did anybody use the ShellscriptAdapter for Solaris/NIS instead - with better result ?
Thanks for reply.

Thank you.
I could solve this problem by using waveset.forceUpdate too.
FYI, I modified Deprovision Fom as below.
      <Field name='accounts[Solaris].waveset.forceUpdate'>
        <Expansion>
          <list>
            <s>delete after action</s>
          </list>
        </Expansion>
      </Field>
      <Field name='resourceAccounts.currentResourceAccounts[Solaris].attributes.delete after action'>
        <Expansion>
          <s>solaris-test-after-delete</s>
        </Expansion>
      </Field>---
n@ruhito

How to get the homedirectory of nis users while they login

HI,
I am having a solaris nis server and a file server running on solaris (pre installed).I need to add new workstations (linux) to this server, i stopped firewall and selinux on linux pc & configured network.I used authconfig to configure nis client. In the NIS server cd /var/yp/src and edited netgroup & hosts.
and then used make command. Now If i give su <nisusername> it allows authentication but without home directory.
Plz tell me which file i need to edit in NIS inorder to get the homedirectory of logged user.
(might be the hostname of the linux system to be given in any file in NIS server)
Thank you All

I found output diffrence to the same command in preconfigured linux client(Redhat 7.3) working well in NIS and the new linux client (RHEL 4.0)
[user@rhelclient ~]$ ps -ef | grep automount | grep -v grep
root 2820 1 0 Jun28 ? 00:00:00 /usr/sbin/automount --timeout=60 /projects yp auto.projects -rw,hard,nosuid,noquota,grpid,intr
root 2833 1 0 Jun28 ? 00:00:00 /usr/sbin/automount --timeout=60 /users yp auto.users -rw,hard,nosuid,noquota,grpid,intr
root 3038 1 0 Jun28 ? 00:00:00 /usr/sbin/automount --timeout=60 /apps yp auto.apps -rw,hard,nosuid,noquota,grpid,intr
redhat7.3:/users/sree> ps -ef | grep automount | grep -v grep
root 1135 1 0 Jun27 ? 00:00:00 /usr/sbin/automount /projects yp
root 1137 1 0 Jun27 ? 00:00:00 /usr/sbin/automount /users yp au
root 1139 1 0 Jun27 ? 00:00:00 /usr/sbin/automount /homes yp au
root 1152 1 0 Jun27 ? 00:00:00 /usr/sbin/automount /eval yp aut
root 1176 1 0 Jun27 ? 00:00:00 /usr/sbin/automount /apps yp aut
/homes and /eval directories are missing in RHEL4.0 and also it have a timeout=60.
Any clues.
thank you

Monitoring Free memory in solaris 8

Hi,
I will like to monitor the amount of free memory in my system. I understand that as Solaris 8 implements "cyclical page cache", the value read from the "free list" column of vmstat will be much more accurate than previous version.
But when i run it on my sparc, the value is 50208KB.
But when i run Memtool 3.8.2, these are the values
Total memory:119Mb
Kernel memory:10Mb
Application:52Mb
Exe and lib :15Mb
File Cache: 33Mb
Free, file cache : 3Mb
Free, free :3 Mb
(33 + 3 + 3) =>39Mb
Can someone tell me how to reconcile the values from this 2 tools? Which is more accurate? (or is there a better method?)
Thanks!

Kalpesh
if you're reading this it would be @&$%#&$( useful to folks in this list to actually BE ABLE TO REFER TO KNOWN SUN BUGS and to be able to report their own.
I couldn't log in to the sunsolve site - do I need a special password to be able to view bug status? (I agree I can find patches there, but
not being able to view/report bugs as a user under the free license program is ridiculous).
How can we get resolution on this?

Help with Solaris.....any help would be nice

I have to write a paper comparing and contrasting Solaris with DOS/MSDOS.....you wouldnt think this would be hard but it is......is ththere a place where I can find a head to head comparsion? If not perhaps some of you familiar with Solaris could give me some points to start with......I should point out I have never used or seen Solaris I am completely new to this area and it may seem simple but the truth is Im a little overwhelmed by all the information....cant see the forrest for the trees...........so any help would be nice ....thanks in advance

Hi,
Solaris is so much more advanced an OS than DOS that a comparison is not straightforward, but here are some less than exhaustive suggestions. Please bear in mind I m not a DOS person, and these are just my personal thoughts on the subject; apologies to Microsoft afficionados for any mistakes.
Similarities:
Both DOS and Solaris are Operating Systems. They provide access to machine resources, like a printer port, or a disk ( through the filesystem), by providing an Application Programming Interface, thus saving programmers from learning the underlying details and machine language of each machine they use. So for example in DOS compiler routines make use of INT 21h to access system services.In solaris we use the system call interface (which does a similar thing underneath).
They also both provide a command shell for ordinary users. In DOS we have command.com, in Solaris you may use Korn shells for example.
Differences
DOS is a single user operating system. Solaris is multiuser, multitasking and multiprocessing.
That is to say Solaris supports multiple programs running on the system at the same time. It will context switch between them to give the illusion of concurrency on a uniprocessor system (time-slicing). On a multiprocessor some tasks will run concurrently.
Many users can log on to a Solaris system (via a serial terminals, or a network connection (telnet or X windows)) and run programs at the same time. Full networking facilities are built in.
Dos only provides the ability to fileshare across the network, using the add on Lan Manager.
Because multiple users run processes (threads/tasks ...) concurrently, Solaris provides extensive security, which DOS doesnt. Each process is owned by a user and runs in its own virtual memory space and is prevented from interfering with any other process memory or other users processes. The operating system and its data is held in protected kernel memory and not directly accessible to processes, so malfunctioning user program cant crash the system accidentally or maliciously.
Devices and files are protected through filesystem permissions, so a device or file has an owner, a group owner and various permissions attributes such as read/write/execute.
Solaris also implements filesystem Access Control Lists (ACLs).
Solaris runs in 32bit protected mode on x86 processors, giving it access to the full address range of the processor (4GB using 4k pages). DOS uses the 16bit segmented memory model.
On ultrasparc systems or the forthcoming Intel Itanium systems, Solaris runs in 64bit mode, giving it access to vastly more memory and
faster computation. (note also DOS only runs on x86 processors).
Who are you writing the paper for? It would probably help some of the readers here with posting suggestions. Some specific questions would be good too.
However, you should try to get access to a solaris system on a network and maybe have a friendly sysadmin show you around a bit.
Do you know anyone with a LINUX system? Solaris and LINUX are variants of the UNIX OS and for your purposes a comparison of LINUX and DOS may suffice.
Hope that is some help.
Ralph
SUN DTS

Can threads be implemented as processes?

Hi all,
a project I'm working on has recently had a JVM added to it. The rest of the
program is a regular single-threaded C written console application. The JVM
has been added for using GUI (swing) components. The program runs on
both Linux and Win32: in the first case it spawns a number of processes,
whilst in the second case it has an equivalent number of threads (circa 16.)
I guess threads can be implemented as processes if the virtual machine
runs in shared memory. But don't understand why it wouldn't use green
threads. Apparently, processes on the Linux box are coordinated via SIGRT_0
and SIGRT_1, after catching bus errors (perhaps null pointers exceptions).
And strace eventually hangs.
Can someone confirm that the processes are actually threads?
Should the main C program be compiled with particular flags?
(Currently we just compile with -dl and dlopen(libjvm.so) at run time
to invoke JNI_CreateJavaVM)
TIA

It is quite old, yes. We have to stick to version 1.3, the actual version is
1.3.1_04 IIRC, because of something coded against specific swing
peculiarities.
Of course threads are more efficient, that's their reason to exist.
However, I haven't been able to find any doc about that, except a single
sentence mentioning that on Solaris the implementation was not using
native threads. BTW, I haven't been able to find the source code for
Linux libjvm either.
Our C application continuously calls into the JVM to get user's input.
Should it use mutexes? Catch signals? I cannot work it out without some
detailed description. Isn't that relevant for programmers? Or is just me
looking in the wrong places?

Mnemonics not working in UNIX

Hi,
Mnemonics set to menus and buttons are working fine in Windows NT, but not in Solaris.
Implementation goes like this:
JButton new = new JButton("New");
new.setMnemonic(KeyEvent.VK_N);
Can anybody help me out in this ??
sku.

hi there, replied to your question, iam sorry i must have not seen it as i was away for a long time from the forum and when i came back i did not have much time to look back at some old posts, hope the solution curbs ur problem, enjoy :)
asrar

SSH / Expired Passwords

I've recently installed a batch of servers with Solaris 10 10/08 and have noticed that the way the Solaris sshd implementation deals with password change on login is different to previous versions of Solaris SSH and/or OpenSSH installed in out environment.
When the user with expired password logs in, he is prompted for a new password. If this password does not meet the complexity standards set for user passwords, the user is then prompted for their original password again instead of being asked to add a valid new password. This has led to a lot of users locking out their accounts because they keep trying to put in the new password
The session output looks like this
ssh -l user serverPassword: <-Enter Existing Password Here
Warning: Your password has expired, please change it now.
New Password: <-Enter new password that does not meet password standards
sshd-kbdint: The password must contain at least 1 uppercase alpha character(s).
Password: <- System requests exising password again
Warning: Your password has expired, please change it now.
New Password: <-Enter new valid password
Re-enter new Password: <-Re-enter new valid password
sshd-kbdint: password successfully changed for user
Any idea why this may be happening?
Thanks.
K

Can you use ssh keys instead? This would allow using your own pass phrase associated with the key you create.
ssh-keygen -t rsa
Now copy the *$HOME/.ssh/id_rsa.pub* file to each site and append the *id_rsa.pub* file to the remote account's *.ssh/authorized_keys* file (repeat the copy and append 29 more times).
Now you should be able to connect based on your ssh key and no longer need to enter the long convoluted password. Instead you just need to enter your own selected ssh key pass phrase.
And you can use *ssh-add* after starting your Mac to add your pass phrase to the ssh-agent already running in the background. Once you do this, ssh will ask the ssh-agent before prompting you for a pass phrase it already knows.
This should totally streamline your ssh and scp access to the 30 remote sites.

Dlopen and JNI

Hi all,
im trying to make a code, that looks like this:
- javaInterface dynamic library: a library that serves as an interface to some java functions
- caller: a library that uses the java interface above and call functions
- executable: calls dlopen to open caller and its functions
JNI crashes while starting the JVM. The interesting thing is, if instead of using dlopen I compile executable linking with caller, it works. I heard there was an incompatibility with Thread Local Storage and dlopen, and the libc RH EL uses has TLS. Can anyone shed a light here?
Thanks
Matheus

It is quite old, yes. We have to stick to version 1.3, the actual version is
1.3.1_04 IIRC, because of something coded against specific swing
peculiarities.
Of course threads are more efficient, that's their reason to exist.
However, I haven't been able to find any doc about that, except a single
sentence mentioning that on Solaris the implementation was not using
native threads. BTW, I haven't been able to find the source code for
Linux libjvm either.
Our C application continuously calls into the JVM to get user's input.
Should it use mutexes? Catch signals? I cannot work it out without some
detailed description. Isn't that relevant for programmers? Or is just me
looking in the wrong places?

Large data plotting

I know this is a bit of a loaded question, but hopefully there is some knowledge that the java gurus can share.
We have a current non-java solaris only implementation of a plotting program that does "Simple point plots". It reads from a file and, it has a calculated axis, colored points, and the ability to zoom. Now the data sets are very large, they can get upwards of 250,000 to 500,000 points with the average being around 30,000 points.
The solaris plotting that we are trying to mimic is XRT. It can load and handle files that huge, and there are no repainting problems, once it is loaded, there is no noticable repainting probelms.
The goal now is do it in java, and have it work on Windows also. We choose JClass, since it is the corresponding java implementation of XRT.
JClass is just to slow. It can load files that big and display the plot, but repainting takes an unuasable amount of time. Maybe even a minute.
The main question is ... is it going to be possible to get the performance we need out of a Java2D implementation, my guess would be no.
If you think it can be done (repaint at least under a second) then what improvement suggestions would you offer.
I you don't than what C based plotting or other solutions could be offered.
Thanks for the time!!
P.S. I will add more dukes, if the answers roll in.

Sure, see below. The only proviso is that I used the latest CVS code, which has changed a bit since the last official release (0.9.13). A new release is expected in the next few days, so if you are not in a hurry I'd wait for that. Alternatively, check out the latest code from CVS at SourceForge:
http://sourceforge.net/projects/jfreechart
Anyway, here is the code for the demo app:
/* ======================================
* JFreeChart : a free Java chart library
* ======================================
* Project Info: http://www.jfree.org/jfreechart/index.html
* Project Lead: David Gilbert ([email protected]);
* (C) Copyright 2000-2003, by Object Refinery Limited and Contributors.
* This library is free software; you can redistribute it and/or modify it under the terms
* of the GNU Lesser General Public License as published by the Free Software Foundation;
* either version 2.1 of the License, or (at your option) any later version.
* This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY;
* without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
* See the GNU Lesser General Public License for more details.
* You should have received a copy of the GNU Lesser General Public License along with this
* library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330,
* Boston, MA 02111-1307, USA.
* FastScatterPlotDemo.java
* (C) Copyright 2002, 2003, by Object Refinery Limited and Contributors.
* Original Author: David Gilbert (for Object Refinery Limited);
* Contributor(s):   -;
* $Id: FastScatterPlotDemo.java,v 1.4 2003/11/12 12:18:07 mungady Exp $
* Changes (from 29-Oct-2002)
* 29-Oct-2002 : Added standard header and Javadocs (DG);
* 12-Nov-2003 : Enabled zooming (DG);
package org.jfree.chart.demo;
import org.jfree.chart.ChartPanel;
import org.jfree.chart.JFreeChart;
import org.jfree.chart.axis.NumberAxis;
import org.jfree.chart.plot.FastScatterPlot;
import org.jfree.ui.ApplicationFrame;
import org.jfree.ui.RefineryUtilities;
* A demo of the fast scatter plot.
* @author David Gilbert
public class FastScatterPlotDemo extends ApplicationFrame {
    /** A constant for the number of items in the sample dataset. */
    private static final int COUNT = 500000;
    /** The data. */
    private float[][] data = new float[2][COUNT];
     * Creates a new fast scatter plot demo.
     * @param title the frame title.
    public FastScatterPlotDemo(String title) {
        super(title);
        populateData();
        NumberAxis domainAxis = new NumberAxis("X");
        domainAxis.setAutoRangeIncludesZero(false);
        NumberAxis rangeAxis = new NumberAxis("Y");
        rangeAxis.setAutoRangeIncludesZero(false);
        FastScatterPlot plot = new FastScatterPlot(data, domainAxis, rangeAxis);
        JFreeChart chart = new JFreeChart("Fast Scatter Plot", plot);
        chart.setLegend(null);
        chart.setAntiAlias(false);
        ChartPanel panel = new ChartPanel(chart, true);
        panel.setPreferredSize(new java.awt.Dimension(500, 270));
        panel.setHorizontalZoom(true);
        panel.setVerticalZoom(true);
        panel.setMinimumDrawHeight(10);
        panel.setMaximumDrawHeight(2000);
        panel.setMinimumDrawWidth(20);
        panel.setMaximumDrawWidth(2000);
        setContentPane(panel);
     * Populates the data array with random values.
    private void populateData() {
        for (int i = 0; i < data[0].length; i++) {
            float x = (float) i + 100000;
            data[0] = x;
data[1][i] = 100000 + (float) Math.random() * COUNT;
* Starting point for the demonstration application.
* @param args ignored.
public static void main(String[] args) {
FastScatterPlotDemo demo = new FastScatterPlotDemo("Fast Scatter Plot Demo");
demo.pack();
RefineryUtilities.centerFrameOnScreen(demo);
demo.setVisible(true);
Regards,
Dave Gilbert
JFreeChart Project Leader

Can you mix calls to the Posix and native thread librariess?

Is it possible to mix calls to different thread libraries within the same process.
We have a shared library interface to our product which is called from a customers application. In order to provide a thread safety we need to protect global data in this library with a mutex. However we do not know which thread library (if any) will be used by the calling application.
If we used a Posix mutex in our library would it work if the calling application was implemented using non-Posix threads?
Are the Posix and Solaris thead header files different interfaces to the same underlying implementation?
What overheads are there loading the thread library when it is not needed? If for example the calling application was implemented without using threads.
Would we have to test our interface with the different thread implementations available? eg /usr/lib/lwp/libthread.so, /usr/lib/sparcv9/libthread.so as well as /usr/lib/libthread.so.

I know the author of the original post wrote this long ago, but I thought I'd write a reply in case anyone else stumbles on this like I just did.
Is it possible to mix calls to different thread
libraries within the same process.Yes. Though you have to be careful about compilation and link flags, and the order of libraries specified on the link command-line (-lpthread and -lthread). For complete information, refer to the Solaris "Multithreaded Programming Guide", available at http://docs.sun.com in the the Solaris 8 (or whatever version) Software Developer Collection.
If we used a Posix mutex in our library would it work
if the calling application was implemented using
non-Posix threads?Yes.
Are the Posix and Solaris thead header files different
interfaces to the same underlying implementation?Yes, I believe so. POSIX specifies an interface. I think the Solaris threads implementation was part of the Solaris OS before pthreads were supported on it. Anyway Solaris really uses "LWPs" (Lightweight Processes) at the kernel level, and user-level threads (POSIX or Solaris threads) are mapped onto LWPs.
What overheads are there loading the thread library
when it is not needed? If for example the calling
application was implemented without using threads.You should never link to the thread library when it is not needed. It results in unnecessary overhead. This is documented in the book mentioned above.
Would we have to test our interface with the different
thread implementations available? eg
/usr/lib/lwp/libthread.so,
/usr/lib/sparcv9/libthread.so as well as
/usr/lib/libthread.so.I don't think so, but judge for yourself; I suspect if any inconsistent behavior is found it will have more to do with defects in these implementations than in your own code. On Solaris 8:
- /usr/lib/libthread.so is the default Solaris threads library that does a many-to-many mapping of user-level threads to kernel-level-lwp's
- /usr/lib/lwp/libthread.so is the "alternate" Solaris threads library that does a 1-to-1 mapping of user-level threads to kernel-level-lwp's
- /usr/lib/sparcv9/libthread.so I think is the fully 64-bit version of the default libthread
Hope this helps somebody.

Ftp fails

The files are in /export/home/username and I want to move them across the network to a new server -- soon to be an LDAP Directory server. I ftp'd from the new server, solaris 10, to the nis/dns/nfs/ftp server, solaris 9.
Would it be best to use ftp or nfs? I have tried both, but both have failed.
What are the possible causes of the ftp failure? Is more information needed? Any suggestions as to how to handle this? Maybe there's a how-to available?
ftp> passive
Passive mode on.
ftp> get printers
227 Entering Passive Mode (nn,nn,nn,nn,216,255)
150 Opening BINARY mode data connection for printers
(1252 bytes).
printers: Operation not applicable
226 Transfer complete.
225 ABOR command successful.
ftp> get solaris-nis-ldap.pdf
227 Entering Passive Mode (nn,nn,nn,nn,164,95)
150 Opening BINARY mode data connection for
solaris-nis-ldap.pdf (51126 bytes).
solaris-nis-ldap.pdf: Operation not applicable
226 Transfer complete.
225 ABOR command successful.
ftp> get openldap-2.3.21-sol10-sparc-local.gz
227 Entering Passive Mode (nn,nn,nn,nn,234,63)
150 Opening BINARY mode data connection for
openldap-2.3.21-sol10-sparc-local.gz (9613279 bytes).
openldap-2.3.21-sol10-sparc-local.gz: Operation not
applicable
426 Transfer aborted. Data connection closed.
226 Abort successful
ftp>

What about "scp"?
scp <file> <user>@<remote host>:<remote path>
Maybe your ftp session fails because you don't have write access to the local directory on your local machine?
.7/M.

AIO persistent connection bug

I am using the solaris kernel implementation of aio (aioread,aiowait). I am having a problem with aiowait once I have four persistent connections to my app. My client app sends data to my server. The server never notifies me that io is ready. If I close one of me connections then the data comes in. I think I may be hitting a hard limit. Anyboby seen something like this?

I am assuming that you are using Solaris aio not POSIX aio.
I think kernel aio is not possible for streams and files. Your aioread/aiowrite may not be going to kernel aio when fd is a socket. Why don't you send the truss output of the program? aioread/aiowrites from/to files will go through pread/pwrite interface but will be done in the context of a separate worker libaio lwp. This would be the case in your aioread/aiowrite (sock). Did you install a SIGIO handler? This is required to get IO done notification.
Follow the notes in man page of aiowait()
NOTES
aiowait() is the only way to dequeue an asynchronous notifi-
cation. It may be used either inside a SIGIO signal handler
or in the main program. One SIGIO signal may represent
several queued events.
--Saurabh Mishra

Solaris 8 NIS Implementation

Similar Messages

Maybe you are looking for