RAID - a very general question concerning stability

Hi everyone,
I've read quite a bit about RAID and the various types of arrays one may set up. Most posts give the following info:
••• RAID 0, while fast, is risky because if one drive dies, everything is lost; good backup routines is essential in this scenario (unless it is only used for scratch disk purposes).
••• RAID 1, and other types of mirrored arrays, are safer because of the built-in duplication of the data on one or more disks.
••• Mirrored arrays, as a general rule, should not be used for backup purposes because, while they duplicate data, they don't replace good backup routines and are therefore susceptible to human error, fires, floods, theft, and general calamity.
But there is one question which I cannot seem to find an answer to:
As I understand it, the stability of any particular array depends on the stability of the disks used. I have used quite a lot of harddrives in my years but have never had a drive fail on me. Perhaps I've been lucky. Then again, I've never hooked drives up using RAID.
My question is therefore:
Using harddrives in any type of RAID setup, will that make them more prone to failure?
Is it a matter of statistical risk? In other words, each drive has a certain, theoretical Mean Time Between Failure and the more drives you have the greater the theoretical MTBF.
Or has it do to with the error handling of the hard-/software controller? I guess it is could be relevant whether you use Enterprise grade drives or not because those, as I understand it, skip errors more easily than non-Enterprise drives and leave the rebuilding of the broken bit up to the controller (of course this is not a RAID 0 scenario).
Basically, do drives become more susceptible to errors because they've been hooked up in a RAID array?
Any thoughts would be very much appreciated.
/pullman

pullman,
No, drives do not become more susceptible to failure as a result of being included in an array.
I'm not sure I totally agree about mirrored arrays not being good for backup purposes. In this regard, one must make the distinction between using a RAID array to make a backup, versus using the array to store a backup. In the first case, the data is stored on the array and the built-in redundancy stands in place of a de facto backup strategy. In the second, the backup strategy includes the use of a RAIDed volume to further safeguard the already-redundant backup (which is made by other means).
In other words, using an external mirrored RAID array to store ones backups seems to me to be an excellent solution. Plus, one can add additional drives to the array, to further increase the redundancy in the system. In a mirrored array, the risk of drive failure is removed from the equation.
However, the risk of a software, or logical failure still remains. By this, I mean potential disk errors. I have suffered far more disk errors than I have actual drive failures. A mirrored RAID array only contains one logical volume, and this volume is just as susceptible to directory failure as any other. However, backup storage on a single mirrored RAID array is no different in this regard as storing it on a single, dedicated drive.
So, we get back to the age-old question: Is a single backup enough? Well, it really comes down to the critical nature of the data being backed up.
Scott

Similar Messages

Sorry - very general question

I have been playing with flash for a while on a Mac but I
have been dissatisfied. It is a pre-intel version and 1) it is very
slow, 2) it does not support Soundforge, which I hear is quite
popular and easy to use. Anyway, I'd like to get a PC but I'm not
sure whether I need a standard computer or a suped up version with
lots of memory, graphics capability, etc. Basically the sites I've
seen have divided computers into those suitable for standard office
tasks and net surfing and those suited for "hard core gaming." I am
basically planning to use the computer just for flash animation and
sound clips. Is a standard PC (that exceeds Flash's minimum
requirements) going to do it for me? Sorry for the general
question. Thanks.

First of all, no need for the apology, there are no stupid
(or general) questions.
Second of all: sorry to all Mac fans, but PCs are the way to
go. After all, you get to choose what company makes your computer.
So you can get either the high-end model from HP or Gateway, or get
something a little more affordable from eMachines or Compaq. With
the Mac, you can only choose, well, a Mac.
While computer shopping, look for the following features:
(This is the BARE minimum of what your expectations should
be)
Windows XP Home Edition
AMD Athlon+ 2.1 GHz Processor
800x600x16-bit or better screen
32 MB Video RAM (For the occasional game here and there)
512+ MB System RAM
Sound Card, Speakers
If you get a computer with Windows Vista (which you probably
will), turn off the Vista visual styles, use Windows Standard.
(Flash has a BIG problem concerning Windows Vista visual styles,
which is the problem I'm having now...see first page of "Flash
General Discussions" for my post, please help! Thank you.)

General questions concerning gaming in (arch)linux

Hi,
I've searched the forums, and I found a lot of howto's that sometimes worked, and sometimes didn't. I'm not a gamer, but have been a longtime linux-user.
I have talked a lot of people into using linux, and almost all of them are happy with there new OS, but my brother (my latest convert) is a gamer 'pure sang'. He's not to happy with the switch to Ubuntu:
* Enemy Territory works, but punkbuster keeps kicking him off the servers
* WoW: he was used to getting rocksolid 80+ FPS in windows, now he gets 50FPS with drops to below 20FPS
* Mumble: giving problems
Yesterday I have found a link to a solution for punkbuster - maybe I can solve that one myself. Is it normal that you have a significant amount lower FPS in linux than in windows? I thought the nvidia-blob was about 'on par' with the windows-driver? I have set WoW to OpenGL and applied several tweaks that I have found around the interwebs...
Should I give archlinux for him a try (I only chose Ubuntu to make it easier for him to update)? Since punkbuster is 32bit-only, should I install 32bit linux for him? Or should I advise him to buy a 'windows 7'-licence key?
PS: Is the Archlinux gaming repo still alive? It looks as if no updates have made since early 2010 and only 1 of the 3 mirrors mentioned in the wiki is still up...
THX!

I am a pretty much a gamer, less so in recent years but... I have found that Linux does not offer the same gaming possibilities as Windows. To this day, I am still dual-booting because of gaming, and I doubt that will change in the near future. I have ran things like Half-life2 and Stracraft 2 in Linux with statisfying results, but with generally fewer frame rates than Windows. As for gaming performance, recent test ran by Phoronix suggest that BSD is faster than Linux, but that's not your question. For myself, I doubt you would find a distro with a significant performance boost for gaming, as they all share the same drivers and all share Xorg to produce graphics. The only reserve is that other test ran by Phoronix suggest that "compositing" as used in Ubuntu's Unity or Gnome 3 has some negative effect on 3D gaming performance, so if you switch to Arch, you might want to consider the "less cute" desktops, like XFCE or LXDE and the like, without compositing. Also, I should mention that using 64 bit Linux to run 32 bit windows games tends to make things a little more difficult. If gaming under Linux with Wine is a must, consider staying under a 32 bit version of the distro of your choice.
As for Windows games, you might want to look at Crossover ( http://www.codeweavers.com/ ) which offers an interface to Wine and better support. There is a "gaming edition" of crossover which might help. Check their DB to see which games are supported, but there's a full-featured demo mode anyway, so try before you buy. Also, you might try playonlinux ( http://www.playonlinux.com/ ) wich offers an interfaces that manages Wine installations for you with "recepies" in the background that are known to offer the best results. If you want the "free" option, playonlinux is the way to go if you want the best chances at running your games, if supported.
Lastly, I would suggest looking at http://www.penguspy.com/ to find Linux native games. Of course, that might not satisfy your brother if he wants specific games.
Last edited by jpsimard (2011-09-27 14:50:16)

A general question concerning .EAR files?

Hi All,
My question is, what should be the file extension of the utility class that
I will include in my ".ear" file? Should it be filename.jar or
filename.war? The team that I am on has successfully been able to deploy a
.ear file that only contains our entity and session beans. We would also
like to "JAR" up our utility class, so that we may be to use the hot
deployment feature of the application server and the removal of the
enviroment variable where the utility classes are located.
We are currently using the Ant utility to help us build our .ear file.
When I give the utility file name a ".war" extension and then deploy the
".ear" file; the application server boots up successfully without throwing
any exceptions. When I give the utility file name a ".jar" extension and
then deploy the ".ear" file, the application server boots up successfully
with exceptions concerning unable to locate one to several utility classes.
Thank you in advance,
Charles
Charles Jean-Baptiste
[email protected]

It sounds like you are having a ClassLoader problem.
The main advantage of deploying EJBs and WebApps in an ear file is that the
ClassLoader for the webapp inherrits from the ClassLoader for the EJBs.
This basicly means that the webapp can see the EJBs...when you put another
util.jar file in the EAR file this will be in a different ClassLoader and
not accesible by either the EJBs or the WebApp
(and visa-versa)....to make this accesible you need to add the entry
Class-Path: util.jar to the manifest file for the EJB.
1)Create a text file newmanifest.txt with a one line entry of
Class-Path: util.jar
2) When you build the original EJB jar file do like this:
jar cmf newmanifest.txt MyEJB.jar com META-INF
-this will create the jar file MyEJB.jar adding directories com and META-INF
and including the contents of newmanifest.txt in the manifest for the jar
file.
3) Build this all into the ear file.
The EJB will now be able to see the util.jar and because the WebApp can see
the same as the EJB it will also be able to see the util.jar.
Gareth
"Charles Jean-Baptiste" <[email protected]> wrote in message
news:[email protected]...
Hi All,
My question is, what should be the file extension of the utility classthat
I will include in my ".ear" file? Should it be filename.jar or
filename.war? The team that I am on has successfully been able to deploya
.ear file that only contains our entity and session beans. We would also
like to "JAR" up our utility class, so that we may be to use the hot
deployment feature of the application server and the removal of the
enviroment variable where the utility classes are located.
We are currently using the Ant utility to help us build our .ear file.
When I give the utility file name a ".war" extension and then deploy the
".ear" file; the application server boots up successfully without throwing
any exceptions. When I give the utility file name a ".jar" extension and
then deploy the ".ear" file, the application server boots up successfully
with exceptions concerning unable to locate one to several utilityclasses.
>
Thank you in advance,
Charles
Charles Jean-Baptiste
[email protected]

VERY general question

This is probably a DNS101 question for most of you, but my problem is this:
I have a computer on my (work) network, that has IP address name resolution working, but I can't seem to find a corresponding A (or ptr record) for it in any of my 3 DNS servers. If I ping the computername, it comes back with a (resolved) IP address,
but no reply. I can RDP to the computer by computername or IP, so I'm guessing ICMP packets are being blocked to it, but my question is....HOW is the network finding the IP, if I can't find a DNS record for it? I know you can alter the hosts file
locally and resolve addresses, but I'm pretty sure that's not the answer, because there are no entries on my hosts file on my client. Thanks in advance.

AHA Robert!
Sorry to all the ninjas who saw my question and thought "well, doy". I came to all this late to the game, after the common usage of WINS and NetBIOS resolution. This computer is in an industrial environment that uses NWlink over
NetBIOS, so I'm pretty sure that was the answer. This has been a good learning experience, thanks all for your kind help.

Signature printing (very general question)

Okay not really an InDesign question, but I know a lot of you know more about printing than I do. I'm strictly on the design end of things:
I always hear about 8-page and 16-page signatures, yet I know as long as my documents are in multiples of 4 I have no trouble getting them printed. What does the printer do when I send him a 76-page magazine? Is it more difficult to print than an 80-page one? I don't know that much about the actual printing though I'm planning to tour a printing plant next week. :-)
Just curious.
Thanks, Phyllis

It's best to talk to the printer because it depends upon the press or presses he is using. It is the size of the press that determines the size of the signatures.
The printer could print four 16 page signatures on one press and three 4 page signatures on another. He could print 80 pages and throw out the waste. He could move the job to a smaller press. His decisions will be based on time and cost. And that will depend upon the size of the press run, the stock, at any number of things. So back to the top, talk to the printer. (And bear in mind that not all printers are created equal. Some are set up for doing odd jobs, and short runs, some aren't.)

... Very general questions

Hi Everyone,
I am a novice/intermediate web designer with experience in
Dreamweaver, but none in Contribute. I have been given an
assignment to build a website for a company, but they would like
control over editing its main content.
How do I define the area where they can edit? How do I give
them the authority to do so? And what do they need to possess to
edit? (they don't actually need to have contribute too, do they?)
Thanks for all your help.
Tara

as a default contribute has 3 user roles for you to use:
1.administrator
2.Publisher
3.Writer
the administrator will be the one setting up the user roles
and permissions
like which page they can edit or who can edit thisparticular
page for security purposes of the site= )
when you download a trial it comes with a manual i suggest
that you read thru deploying contribute or go thru the help
files,that will get you running in no time=)

General questions concerning RAC installation

Dear Oracle experts,
I would like to try an RAC installation using Oracle 10g on Solaris(Sparc).
I just want to go through the installation process without the intention to use
the RAC productively.
Is it possible to install the RAC database in a second ORACLE_HOME additionally to
an existing Oracle installation ?
Are there any traps to avoid ?
Do I really have to install the database software after installing the Cluster software or
is it possible to install the cluster software to upgrade an existing database installation ?
Is it possible to use the RAC without ASM ? (My second machine doesn't have a harddisk left..).
Thanks in advance for your advices.
Best regards,
Daniel

Hi Daniel,
You can't have only one instance of ASM running and sharing the disks. You need to have ASM installed on each RAC node in order to mount the shared disks.
Now you can use one single ASM instance on a server that could be shared by multiple databases (on that same server). We do this in one of our development environments.
You may not have to upgrade your hardware for a little RAC sandbox. The main thing is getting storage shared between the 2 servers and having the extra NICs for the private interconnect.
Alfredo

A very basic question regarding data block

Hi All,
I've a very basic question concerning data blocks in oracle forms 10g.
I want to make a view only screen (only query allowed, no update, insert or delete).
I'll have 6-7 fields on the screen but all the fields are not from a single table.
For e.g, let say we've field names to display on the screen are f1, f2, f3, f4..
Out of this f1 and f2 will come from table A and f3, f4 will come from table B.
Now, my question : Is it possible to create a data block using the data block wizard for such situation if we select create data block from table options?
If no, can you please tell me an approach to do this.
Regards,
Navnit

Hello ,
First write your query & select datablock property.
just change the below properties
Query Data Source Type=From Clause Query
Query Data Source Name = (Paste query here)
now you should add the block ITEMs and give their names according to query columns names. shows the column on canvas and run..
Best Regard
skyniazi
Edited by: SKYNIAZI on Mar 29, 2009 1:32 PM

Very basic questions about hard drives

I'm upgrading from vintage Beige G3, so much of the jargon about internal drives is new to me.
My MacPro will come with an OEM 160 GB drive. I will use this for my primary boot drive -- OS and Apps. All my data, photo files etc. will be kept on my second/third drives. I do not plan any sort of RAID setup. Windows is going nowhere near my Mac.
For the extra drives --
1) SATA? SATA II? What's the difference and is it critical?
1a) What is a "raptor" drive?
2) Does size matter (we're talking about GBs here, guys...)
3) Reliability is more important to me than speed or noise. We're a long long way from any technical help or even UPS.
4) Do new drives need to be formatted before using them internally?
5) If I partition one of the extra drives so I can keep a clean, backup OS boot partition, I need to format it as GUID, correct? Or does that just apply to external firewire boot drives? Does it matter which partition will be the boot partition?
6) I've read discussions about firmware requirements for some drives. What's that all about?
I was leaning toward MaxLine Pro 500 GB drives, but have been reading (reading too much, perhaps?) and now have all the above questions/concerns. (Whatever happened to the good ol' days of "buy an IDE drive, insert, play"?)
Thanks for your patience with these questions. I need to get it right the first time because of our remoteness!

1. SATA 1 interface speed rated at 1.5 Gb/s; SATA II interface speed rated at 3.0 Gb/s.
2. Drive ratings/performance is determined by many factors. You can best find out about individual drive performance by visiting http://www.storagereview.com/ to see their benchmark results on various drives. You will also find some helpful comparisons at http://www.barefeats.com/.
3. As for drive specs see previous answer. Generally speaking within any manufacturer's category of drive models for a given size the fastest drive will be the most expensive. Manufacturer specs usually do not provide any idea about drive performance in real world situations. Only benchmark comparisons are really useful, and even those can be confusing.
4. Yes, you can boot an Intel Mac from an APM partitioned drive. You simply cannot install OS X from the installer disc because it will refuse to install on a drive that has not been GUID partitioned.
5. Some older Seagate 750 GB models were not working properly when installed in a RAID configuration on Mac Pro computers. Seagate provided a firmware update to fix the problem for those who already owned the affected drives.
I don't know where you got the idea that Intel Macs were "picky about such things." Nor that installing drives in a beige G3 was a no-brainer. The old beige G3 required a lot of mechanical effort to mount a new drive internally. You not only had to connect all the various cables, but if you had two drives on the same ribbon cable each drive had to be jumpered properly - one as MASTER and one as SLAVE. Some older ATA hard drives had two different jumper settings for a MASTER - MASTER solo and MASTER w/SLAVE. Some beige G3 models wouldn't even boot OS X unless it was installed on a MASTER drive. And, you could not install OS X on a large drive except by partitioning the drive with an 8 GB partition for OS X that had to be the first partition on the drive. RAIDs may work or not depending on if you ran OS 9 or OS X. At least to me that wasn't the "good old days" unless you were comparing installing ATA drives to dealing with SCSI. I would say that in your particular case you were lucky, and you never tried putting two drives on the same ribbon cable. Oh, and I forgot about the size limiting jumper on most ATA large drives to limit capacity to 32 GBs because that's all that was supported by FAT32. Oh, yes, and the beige G3's disk controller did not support drives larger than 128 GBs.

Report Builder Question - OA AR Aging - and a general question

I'm sure this is the wrong forum for this question, but I thought there might be someone here who might be using Oracle Applications and Report Builder who'd be kind enough to help me out.
We've recently implemented Oracle Applications 11.5.10 and I have to use report builder to change the Accounts Receiveable Aging (7 bucket) to a 5 bucket report. I've already made some changes to the seeded "ARXAGMW.rdf" report, but I'm not a big Oracle Reports guy. I've stumbled through making some changes in various other reports. But this one is just plain nasty! :)
I was thinking that I could simply add buckets 6 & 7 to bucket 5, then just hide or delete the 6 & 7 buckets. But I'm not sure where to even start. Any help with this would GUARANTEE a Christmas or other holiday card this year! :)
I really want to keep this simple as possible, so any help would be very....helpful. :)
Oh, my general question is: Are there any resouces/books for Oracle Reports (Report Builder)? I feel so lost trying to modify existing reports, let alone creating new ones.
Thanks again!
Steve

Hi Steve,
I am working on the 7-bucket aging report and i want to add a new field in data model.
As the query is build dynamically, i have modified the function BUILD_CUSTOMER_SELECT to meet my requirements.
But the problem is that in the data model, the field is not present in my Grouping. and if I try to add the field in the Data Model query (Q_ Customer) section,
i get the following error: ORA-01789: query block has incorrect number of result columns.
The query is as shown below:
select rpad('a',50,'-') short_cust_name,
0 cust_id,
rpad('a',30,'-') cust_no,
rpad('a',500,'-') sort_field1,
rpad('a',40,'-') sort_field2,
0 payment_sched_id,
rpad('a',32,'-') class,
sysdate due_date,
0 amt_due_remaining,
0 days_past_due ,
0 amount_adjusted,
0 amount_applied,
0 amount_credited,
sysdate gl_date,
'x' data_converted,
0 ps_exchange_rate,
0 b0,
0 b1,
0 b2,
0 b3,
0 b4,
0 b5,
0 b6,
rpad('a',25,'-') bal_segment_value,
rpad('a',500,'-') inv_tid,
rpad('a',32,'-') invoice_type
, 'y' parent_cust --I WANT A NEW FIELD HERE TO BE VIEWED ON THE REPORT LAYOUT LATER
from dual
where 1=2
UNION ALL
&common_query_cus
Did i missed somthing 4 me to be able to add the field here?

Good afternoon ladies and gentlemen! My question concerns the impossibility to open RAW-files directly from the program Adobe Bridge. At the moment when you open a RAW-file from Adobe Bridge by double-clicking, RAW-file is opened only in Photoshop. In t

Good afternoon ladies and gentlemen!
My question concerns the impossibility to open RAW-files directly from the program Adobe Bridge. At the moment when you open a RAW-file from Adobe Bridge by double-clicking, RAW-file is opened only in Photoshop. In the settings Adobe Bridge - in "open RAW-files by double-clicking in Adobe Camera Raw» box is checked. When you try any changes in the settings Adobe Bridge system displays a message:
Bridge's parent application is not active. Bridge requires that a qualifying product has been launched at least once to enable this feature.
The entire line of Adobe products on my computer updated to the latest updates. Previously, a family of products Adobe Photoshop on your computer is not set. Computer - PC, Windows 7 Enterprises.

<moved from Adobe Creative Cloud to Bridge General Discussion>

Camileo charging problem (solved) and a general question!

Hi all!
First of all, I was going to ask for help as to why the Camileo S10 was not charging (the orange light wasn't flashing), and I'd seen a few people with similar problems.
The solution?
Give the contacts on the battery a clean.
The insulation sticker that comes on it must leave some kind of residue on it, and it's enough to prevent charging. Now it's flashing away happily :]
So my general question was, is it possible/advisable to use the camera on the mains?
Rather than constantly draining and charging the battery during long shoots, I'd prefer to just leave it plugged in!
Thanks very much in advance!
Paul

Hi
I think the battery handling is always the same no matter what product it is
From time to time the battery should be recalibrated.
This means that the battery should be discharged fully and after then you should charge it again until the battery would reach 100%
I do this with all my batteries; mobile phone battery, digi cam battery and notebook battery.

Update on Bioinformatics WIKI, scripting challenges, and a general question

I am waiting for my site to go on-line at Oak Ridge National Labs (USA, Tennessee). Should be another week or so, maybe less.
When that happens, you will see a veritable explosion of scripting challenges in my wiki (Emerging Technologies->Bioinformatics.)
One general question in preparation for these challenges.
There are a number of standard bioinformatic programs that can be run interactively via the web at various sites, e.g. "BLAST" and "STRIDE".
Although these can also be run locally, this requires that you download large databases and keep them updated.
So here's my question to the scripting experts:
Are scripting languages powerful enough to submit queries to web pages and then use regex's to parse the html that is returned?
Bill Mann has used PERL to do some of the required regex parsing, but there is a lot left to do, and, his stuff only works when a perl program is invoking a bioinformatic program locally, not interactively.
If so, we all can do some beautiful stuff together , if anyone is interested ...

...my wiki on...
There is by it's very nature no such thing as MY WIKI, except you run your own wiki project in an exclusive mode. Which were...well ... unusual.
Are scripting languages powerful enough to submit queries to web pages and then use regex's to parse the html that is returned?
Yes.
anton

Some general questions about multi-threading

Hey Everybody,
I have a dilemma with a multi-threaded program that I have written.
Well to be truthful I actually wrote the program with no consideration whatsoever for multithreading.
As a result of this complete lack of concern I get a completely hung user interface.
My program performs a lot of network communications over URL��s and also writes information from URL��s to disk. The combination of networking and I/O in my program and the fact that I have not built multi-threading into the program has lead to complete failure of the user interface. When I say complete failure I mean complete failure. If it were only button��s not responding then I wouldn��t be all that worried, however the entire drawing area of my programs ��Frame�� completely freezes. It��s cool if nothing is maximized or minimized over the frame but as soon as this happens my programs drawing area has a ��serious hang over!��.
I have begun learning about threads, in principle they sound wicked however in practice they have proven to be a brain tease (oops �K.. honesty).
The difficulty of threads should not be all that daunting to me, however I have a couple of very simple questions that I would love to have answered�K..anybody out there!!!!!!!!!
This is the normal basic method of running a thread that I have been attempting to incorporate into my program:
1). Extend the thread class,
2). Override the run() method in the subclass (created from step 1),
3). Create an instance of the subclass,
4). Call the start() method on the instance (from step 3).
I have read that every statement the thread will execute is contained within it��s run() method.
Question 1). This being the case is it possible to have member variables or methods inside a class that extends thread? Please explain!
Question 2). Can a class that extends thread contain a constructor? Please explain!
Question 3). If a subclass of thread cannot contain a constructor as I assume to be the case then how can code executing within the newly spawned thread reference information from the object that spawned it?
If anybody out there on the great net can answer even half of one of these questions I will be more than grateful. Thanks for your time, and rock on Java!
David

Thread t = new Thread(MyRunnableDerivedClass);
t.start();I hope it is understood that MyRunnableDerivedClass is actually an object of the MyRunnableDerivedClass. sorry for the confusion.
so here's a thread that reads from a file using constructors, member variables, other funcs. It's not optimal but shows use of all of the above. I just typed this in so there are probably syntax errors, but you should get the point.
public class ThreadedFileReader implements Runnable
    private File m_File;
    byte[] contents = new byte[0];
    public ThreadedFileReader( File f )
       m_File = f;
    public void run()
        BufferedInputStream bis = null;
        try
            bis = new BufferedInputStream( new FileInputStream( m_File ));
            int avail = bis.available();
            while (avail > 0)
                int oldLen = contents.length();
                contents = expand( contents, avail );
                bis.read( contents, oldLen, avail );
                avail = bis.available();
        catch (Exception e)
        { //do something witty
        finally
            try{ if (bis != null) bis.close(); } catch (Exception e){}
   private byte[] expand( byte[] oldBuf, int addedLen )
      byte[] newBuf = new char[oldBuf.length + addedLen];
      System.arrayCopy( newBuf, 0, oldBuf, 0, oldBuf.length );
      return newBuf;
   public byte[] getContents()
       return contents;
//here's where we use it
File f = new File( "c:\myfile" );
ThreadedFileReader tfr = new ThreadedFileReader( f );
Thread t = new Thread( tfr );
t.start();
t.join();
// At this point the thread has died, but the thread object still exists
byte [] data = tfr.getContents();

RAID - a very general question concerning stability

Similar Messages

Maybe you are looking for