Corrupt jar message

36 views
Skip to first unread message

Patricia Shanahan

unread,
Aug 5, 2014, 9:36:55 PM8/5/14
to rio-...@googlegroups.com
I downloaded
http://www.rio-project.org/releases/rio-5.0-M3-installer.jar, and tried
to run it. I get a JVM dialog "Error: Invalid or corrupt jarfile".

I've repeated the download, in case it was a failure to download the
entire JAR, with no change.

Do I need a specific JRE version?

Thanks,

Patricia

Dennis Reedy

unread,
Aug 6, 2014, 6:35:40 AM8/6/14
to rio-...@googlegroups.com
Hi Patricia,

I'm on travel right now with limited connectivity, and will look into this when I get back. What version of Java were you using?

Thanks

Dennis

Sent from my iPhone
> --
> You received this message because you are subscribed to the Google Groups "Rio Users Group" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to rio-users+...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

Dennis Reedy

unread,
Aug 6, 2014, 1:49:37 PM8/6/14
to rio-...@googlegroups.com
Hi Patricia,

Just gave this a try and had no issue. The machine I got access to had Java 1.6, I can try with 1.7 and Java 8 when I get back home. What JVM are you using?

Dennis

Patricia Shanahan

unread,
Aug 6, 2014, 2:09:30 PM8/6/14
to rio-...@googlegroups.com
Thanks for trying that. Unfortunately, just after starting this I got
really busy with some personal stuff, so I may not be as responsive as I
would normally be after asking a question.

Patricia Shanahan

unread,
Aug 6, 2014, 6:14:46 PM8/6/14
to rio-...@googlegroups.com
I deleted all Java installs and just installed the last Java 1.6. It
still fails, so my original Java version guess is wrong.

Do you have an MD5 sum check for the installer jar?

Patricia

On 8/6/2014 10:49 AM, Dennis Reedy wrote:

Dennis Reedy

unread,
Aug 6, 2014, 6:57:17 PM8/6/14
to rio-...@googlegroups.com
Back to limited connectivity again. Can anyone else on the list try and download this?

Thanks

Dennis

Sent from my iPhone

John Sarman

unread,
Aug 6, 2014, 7:24:30 PM8/6/14
to rio-...@googlegroups.com
Is it possible that a jar in your maven .m2 folder is corrupted? You can change settings.xml to change you maven repo home temporarily and do a clean build to test. This will force maven to redownload all dependencies.  

John Sarman 

John Sarman

unread,
Aug 6, 2014, 7:35:55 PM8/6/14
to rio-...@googlegroups.com
Just wanted to add the details to the vague test.

in your .m2 folder (typically $user.home/.m2)  either add or edit the settings.xml file

<settings>
<localRepository>${user.home}/.m2/repotest</localRepository>
</settings>

if you have an existing settings then just add the localRepository line.  


This just tells maven that the repo folder (default)  is called repotest.  When you do a build maven will not be able to find the jars and it will trigger a download of the dependent jars.

If you happen to run a maven mirror locally such as sonatype, then this may not prove anything if the mirror has the corrupt jar file, because you'll just be re-downloading the potential corrupted dependency. 

  




Dennis Reedy

unread,
Aug 6, 2014, 7:36:57 PM8/6/14
to rio-...@googlegroups.com
Hi John,

Thanks for looking into this. Its not a Maven issue, its a direct download (the installer) from the Rio site. I can grab it okay, looks like Patricia has issues with the jar.

Thanks

Dennis


On Wed, Aug 6, 2014 at 7:24 PM, John Sarman <johns...@gmail.com> wrote:

John Sarman

unread,
Aug 6, 2014, 7:40:51 PM8/6/14
to rio-...@googlegroups.com
In that case here is the md5 for Rio 5.0 M3.       36eacf3557921dc07fec2ee29c8990b0

John Sarman

unread,
Aug 6, 2014, 7:44:58 PM8/6/14
to rio-...@googlegroups.com
Sorry that MD5 is for the zip not installer

Dennis Reedy

unread,
Aug 6, 2014, 7:47:53 PM8/6/14
to rio-...@googlegroups.com
Hi John,

Yep, but it's the installer, not the standard distribution. If you go to the http://rio-project.org/download.html page, you'll see a link for "Download Rio 5.0 M3 using the installer". It's that jar thats is an issue for Patricia. I can download and run it without issue (could you check?). Of course in my infinite wisdom when I published the installer for 5.0-M3, I did not add an MD5 checksum (which of course I do for the regular downloads, just not for the installer for some reason :/).

Thanks

Dennis

John Sarman

unread,
Aug 6, 2014, 7:50:11 PM8/6/14
to rio-...@googlegroups.com
mac:tools jsarman$ md5 rio-5.0-M3-installer.jar
MD5 (rio-5.0-M3-installer.jar) = fa35c945e3ec9689c04f7243b298b2df

John Sarman

unread,
Aug 6, 2014, 7:51:27 PM8/6/14
to rio-...@googlegroups.com
also I launched it successfully on win7 with java -jar rio-5.0-M3-installer.jar

Dennis Reedy

unread,
Aug 6, 2014, 8:05:13 PM8/6/14
to rio-...@googlegroups.com
John,

Thanks for taking the time.

Patricia,

Does any of this help?

Regards

Dennis

Patricia Shanahan

unread,
Aug 7, 2014, 12:44:03 AM8/7/14
to rio-...@googlegroups.com
$ md5sum *.jar
2162020d4ce84024e923a4d7161613a0 *rio-5.0-M3-installer.jar

It looks as though the problem may be download related, rather than
environment or JRE. I may be able to take another look at this tomorrow
afternoon.

Thanks,

Patricia

Patricia Shanahan

unread,
Aug 7, 2014, 9:33:57 PM8/7/14
to rio-...@googlegroups.com
The combination of successful runs with Java 1.6, and a checksum
mismatch, suggests that it may be a file downloading problem. That will
be my next approach.

Patricia Shanahan

unread,
Aug 12, 2014, 4:31:41 PM8/12/14
to rio-...@googlegroups.com
The problem is definitely download related. I've have been using
Firefox, but I have also tried Chrome in case it was a browser issue.

Each time I download, the resulting file is a different length. I've
used "cmp" on some pairs of files. The first difference is always EOF on
the shorter file. It appears that my problem is premature termination of
the downloads.

I have downloaded many other files using Firefox without any problems -
that's why I looked first for Java version issues.

I have been doing these downloads on a Windows 8.1 system.

Here is some sample cygwin output, showing file lengths and cmp results.

=======================================================================

Patricia@Jan2014Desktop /cygdrive/c/Documents and
Settings/Patricia/Downloads/Rio
$ cmp bad2_rio-*.jar rio*.jar
cmp: EOF on rio-5.0-M3-installer.jar

Patricia@Jan2014Desktop /cygdrive/c/Documents and
Settings/Patricia/Downloads/Rio
$ cmp bad2_rio-*.jar bad3_rio*.jar
cmp: EOF on bad3_rio-5.0-M3-installer.jar

Patricia@Jan2014Desktop /cygdrive/c/Documents and
Settings/Patricia/Downloads/Rio
$ ls -l *.jar
-rwxrwx---+ 1 Patricia None 22460695 Aug 5 18:14
bad_rio-5.0-M3-installer.jar
-rwxrwx---+ 1 Patricia None 20483071 Aug 8 16:46
bad2_rio-5.0-M3-installer.jar
-rwxrwx---+ 1 Patricia None 16016719 Aug 12 13:15
bad3_rio-5.0-M3-installer.jar
-rwxrwx---+ 1 Patricia None 18316687 Aug 12 13:21 rio-5.0-M3-installer.jar

Patricia@Jan2014Desktop /cygdrive/c/Documents and
Settings/Patricia/Downloads/Rio
$

=======================================================================

Patricia

Wade Chandler

unread,
Aug 12, 2014, 8:01:47 PM8/12/14
to rio-...@googlegroups.com

Are you able to open the file in an archive browser such as WinZip or izarc? I can download and extract the archive on my Android phone and open files inside it.

The MD5 checksum of the file is fa35c945e3ec9689c04f7243b298b2df for me. The sha1 is 21a5ddd9a9948649664077fa4a89fa20fede52a5. Does either of those match for you?

Do you happen to use a proxy server?  Could be that. Could be your ISP. On a project I was working on, one of our clients was not able to download certain size files correctly through our application. It turned out to be a problem with their ISP. They were truncating and corrupting certain size files; once they were over a certain size this occurred.

Wade

--
You received this message because you are subscribed to the Google Groups "Rio Users Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rio-users+unsubscribe@googlegroups.com.

Wade Chandler

unread,
Aug 12, 2014, 8:06:05 PM8/12/14
to rio-...@googlegroups.com

I did not notice this before, but mine and John's checksums match. Sorry for posting the obvious question. I really believe something is just wrong with your download or network connection. Are you able to download other files over say 20 megabytes in size and open them correctly?

Wade

Patricia Shanahan

unread,
Aug 12, 2014, 8:51:20 PM8/12/14
to rio-...@googlegroups.com
The problem is definitely a truncated file. Each time I attempt a
download, it gets a different size and md5 checksum, none of which match
the checksum reported by people who can run the JAR. The first
difference reported by cmp on a pair of my download results is always
EOF on the shorter file.

In between a couple of these attempts, I downloaded and ran
jdk-7u67-windows-x64.exe, which is over 132 megabytes. No problem.

I do go through a proxy, so I'll try disabling that.

Thanks for your interest, suggestions, and confirmation of the correct
checksums.

Patricia
>>> email to rio-users+...@googlegroups.com.

John Sarman

unread,
Aug 12, 2014, 9:08:24 PM8/12/14
to rio-...@googlegroups.com
Patricia,
Sorry you are having a one off problem. I mean I am guessing no one else is experiencing this issue, but your setup is truly causing the jar corruption on download. Have you tried another computer to confirm that the issue is not your specific os( meaning hardware plus installed os).
At this point if I were in your situation, I would have convinced myself that my os was corrupted with a seriously bad virus, that was doing a horrible job at rewriting a jar file, to further corrupt my os. To which , I would be in the process of a clean install of a new os. Of course this is just my opinion.

Still I feel it weird and even weirder that you receive a set of 1 s and 0 s that differ from everyone else that attempts to reproduce your scenario.

Good luck in resolving the issue and I am extremely curious on your results!

John S.

Patricia Shanahan

unread,
Aug 12, 2014, 9:08:46 PM8/12/14
to rio-...@googlegroups.com
Disabling the proxy made no difference - yet another truncated JAR file,
with yet another incorrect checksum. :-(

My ISP is Earthlink. I've used them for many years, and this is the only
file for which I've had this sort of trouble.

I tried downloading the Rio 5.0 M3 zip file, and it was fine - correct
length and MD5 checksum, and unzips without error. My download problem
seems to be very specific to that one installer file.

Patricia

Patricia Shanahan

unread,
Aug 12, 2014, 9:50:44 PM8/12/14
to rio-...@googlegroups.com
I don't think I'm getting different 1s and 0s - just too few of them.

This is the only file that seems to be affected. A virus would be far
more likely to prevent updates to AVG or Windows, or installing a
different version of Java, and those things are working fine.

Of course, there is a first time for everything, but I am careful about
the sites I visit and the software I install, and I read e-mail in
text-only mode. So far have never needed to resort to an OS install to
remove a virus.

I seem to have the same problem on my Windows 7 laptop, so it is not OS
specific.

I did get a new DSL modem a few weeks ago, but I've done several large
downloads since then.

Patricia

John Sarman

unread,
Aug 12, 2014, 10:07:36 PM8/12/14
to rio-...@googlegroups.com
So basically most unconventional reply I would ever give, but how about setting up TOR then download through TOR network to either prove, or disprove that earthlink is causing the issue. Basically if you download through a TOR connection then earthlink cannot discriminate on the jar file, because from the data you are transceiving will just be encrypted 1 s and 0 s from earthlinks view of the data.  The binary data from their point of view is just an encrypted stream.  Again just an idea to find the root cause of your downloading issue. Or even better if you have ability to use a VPN to get the jar do that. In either scenario you are using an encrypted channel to download content.  Again this is just a thought from a debuggers point of view to rule out your subscriber.     



-- You received this message because you are subscribed to the
Google Groups "Rio Users Group" group. To unsubscribe from this
group and stop receiving emails from it, send an email to

--
You received this message because you are subscribed to the Google Groups "Rio Users Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rio-users+unsubscribe@googlegroups.com.

Wade Chandler

unread,
Aug 13, 2014, 9:35:13 AM8/13/14
to rio-...@googlegroups.com

Do you have an Android phone? If so could you download it over your mobile internet connection? If you go to the Google Play Store, and search for a checksum program, you will find one you can use to at least calculate the checksum. Being sure that you have done this over your mobile connection versus your local WiFi, generate a checksum of the file. Confirm with the one we have sent you. Next, connect your phone to your WiFi connection . Then download the file again. Produce the checksum again. Compare the two.

Hopefully, given your geographic location, you will be routed through most of the same server infrastructure on the Rio side. Then, given the results of that, we can try a couple other things to hopefully narrow down where the problem is.

Wade


For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "Rio Users Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rio-users+unsubscribe@googlegroups.com.

Patricia Shanahan

unread,
Aug 13, 2014, 10:38:45 AM8/13/14
to rio-...@googlegroups.com
iPhone only. I do have an Android tablet, so one possibility would be to
take it to a WiFi hotspot and try the download there. Unfortunately, I'm
also dealing with some urgent personal business, so I won't get to do
that immediately.

Meanwhile, perhaps someone could check my theory that the problem is
only truncation, not wrong data. The following output shows the sizes
and MD5 sums for some of my downloads:

Patricia@Jan2014Desktop /cygdrive/c/Documents and
Settings/Patricia/Downloads/Rio
$ ls -l *.jar
-rwxrwx---+ 1 Patricia None 22460695 Aug 5 18:14
bad_rio-5.0-M3-installer.jar
-rwxrwx---+ 1 Patricia None 20483071 Aug 8 16:46
bad2_rio-5.0-M3-installer.jar
-rwxrwx---+ 1 Patricia None 16016719 Aug 12 13:15
bad3_rio-5.0-M3-installer.jar
-rwxrwx---+ 1 Patricia None 18316687 Aug 12 13:21
bad4_rio-5.0-M3-installer.jar
-rwxrwx---+ 1 Patricia None 22559431 Aug 13 03:06 rio-5.0-M3-installer.jar

Patricia@Jan2014Desktop /cygdrive/c/Documents and
Settings/Patricia/Downloads/Rio
$ md5sum *.jar
2162020d4ce84024e923a4d7161613a0 *bad_rio-5.0-M3-installer.jar
d9bfdca25b7f2f1072bc255481d54b07 *bad2_rio-5.0-M3-installer.jar
cabe99b861335e0ef1f6c163461734fd *bad3_rio-5.0-M3-installer.jar
11cf0d42f84735cc1093330dbee1433c *bad4_rio-5.0-M3-installer.jar
c32d90729aa55d4bb4b014bfe47237bc *rio-5.0-M3-installer.jar

If the theory is correct, truncating a correct download to the length of
one of my files will give a file with the matching sum check.

Thanks,

Patricia
>>>>>> email to rio-users+...@googlegroups.com.
>>>>>> For more options, visit https://groups.google.com/d/optout.
>>>>>>
>>>>>>
>>>>>
>>>>
>>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "Rio Users Group" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to rio-users+...@googlegroups.com.

Wade Chandler

unread,
Aug 13, 2014, 12:17:52 PM8/13/14
to rio-...@googlegroups.com

I will try out truncating sometime later today for you just to be sure, but you should be able to download the file with your iPhone and find a free checksum tool on it. Then you can verify more things that way. One of the next steps I was interested in seeing if you get different results between your connections, mobile vs WiFi, was if you hooked your computer directly to your DSL modem if that worked. Essentially trying to see if it is only an ISP issue or possible hardware.

If you get similar results, then I might have to look to see if there is some cache flushing which could be done on the Rio server side. Since John, Dennis, and I seem to have successfully downloaded it from various geographic locations I am doubtful that is the issue, but you never know until you know.

Wade


For more options, visit https://groups.google.com/d/optout.


--
You received this message because you are subscribed to the Google Groups
"Rio Users Group" group.
To unsubscribe from this group and stop receiving emails from it, send an

For more options, visit https://groups.google.com/d/optout.


--
You received this message because you are subscribed to the Google Groups "Rio Users Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rio-users+unsubscribe@googlegroups.com.

Patricia Shanahan

unread,
Aug 13, 2014, 12:22:18 PM8/13/14
to rio-...@googlegroups.com
Most of my download attempts were on a computer with a wire ethernet connection to the modem, not WiFi. I'll look into using my phone. 

Patricia
To unsubscribe from this group and stop receiving emails from it, send an email to rio-users+...@googlegroups.com.

Patricia Shanahan

unread,
Aug 16, 2014, 9:53:00 PM8/16/14
to rio-...@googlegroups.com
The problem has gone away!

I have been very busy for a few days, so I didn't make any attempts.
Today I tried the download again. The file was full size and a valid
JAR. To check whether it was a one-off, I downloaded the file again, and
got an identical file.

I have no idea what has changed.

Patricia

Dennis Reedy

unread,
Aug 17, 2014, 5:35:06 AM8/17/14
to rio-...@googlegroups.com
Glad to hear it!

Sent from my iPhone
Reply all
Reply to author
Forward
0 new messages