I've checked all the files both directly on the server and after downloading them via the web portal from the server to my workstation, and they all seem to check out. The "tar -tf ..." command also works as is should on all of them.
Can you verify that the file you downloaded matches the md5sum, sha1sum, and/or gpg signature we have listed? You can see the sums and signatures by clicking on the 
icon next to the file names.
Thanks.
Hi Daniel,
So I tried to download the file with wget, and what I got works with tar -tf.
Then I tried to download with Chrome instead of Firefox, and that also works.
So it's only Firefox that does not work.
A web search turned up this conversation about the same problem:
Now if I gunzip the firefox downloaded foo.tar.gz file, it gives me a foo.tar file, and if I then rename this from foo.tar to foo.tar.gz, now tar -tf works! So the file is somehow getting zipped twice.
So next I decided to investigate the http headers your server is sending. Here they are for the working and non-working downloads:
Connection: keep-alive
Content-Disposition: attachment; filename=mariadb-java-client-1.4.6.jar
Content-Encoding: gzip
Content-Type: text/html; charset=utf-8
Date: Mon, 10 Oct 2016 19:55:03 GMT
Etag: W/"575f35fe-5ac03"
Last-Modified: Mon, 13 Jun 2016 22:38:54 GMT
Server: nginx
Transfer-Encoding: chunked
Connection: keep-alive
Content-Disposition: attachment; filename=mariadb-connector-odbc-2.0.12-ga-linux-x86_64.tar.gz
Content-Encoding: gzip
Content-Type: text/html; charset=utf-8
Date: Mon, 10 Oct 2016 20:26:34 GMT
Etag: W/"57dac91c-a66b4"
Last-Modified: Thu, 15 Sep 2016 16:15:24 GMT
Server: nginx
Transfer-Encoding: chunked
"Content-Encoding: gzip" says the file transmitted over the Internet will get zipped, and the browser should unzip it when it receives it.
So I'm guessing then that the tar.gz source file on your end, is getting redundantly zipped by your web server, i.e. it's a tar.gz file in your server folder, and your servers is set to also zip the file, thus zipping it twice.
Then Firefox takes a look at what it got, with the tar.gz extension, guesses wrong, and does not unzip it, assuming it is a server error to double zip it.
Even though wget and Chrome do unzip this, I think it's nevertheless an error to be double zipping it on your server. For files that are already zipped the http server header should not be "Content-Encoding: gzip".
What do you think? Does this make sense to you?
In any event, Firefox is a common enough browser that it should work without fussing with it's internal settings. And I have not had this trouble downloading other tar.gz files from other sites. So I think a small change to your server will make this problem go away.
For me, for now I have a workaround.
Thanks for your help.
– Howard
Could you please try to gunzip it first, and then try the same tar command? Worked for me and xf, too. I am not sure why that can happen.
Upd. Sorry, did not read previous comments. I also downloaded with FF. And it apparently looks like has been gzipped twice
BTW, for completeness, my Firefix is: 45.4.0 Firefox-ESR (Extended support, meaning the version is more stable over time than with the normal Firefox).
I the reasons I use Firefox and not chrome, because I have a dual boot system w/ Debian/Windows10, and 1) Firefox will use the same profile for both OS, and so when I switch from on OS to the other, my browser tabs are the same on both operating system, and 2) the version does not change as often, lessening the chance of messing up the profile that is shared by both systems.
@Novitsky, Hi, I had already tried that, but after the gunzip, the tar x would fail. Please be sure to read my extended comments 2 back 20 minutes ago. Thanks!
sorry, sorry. I did not reload the page before commenting. and somehow I received mail notifications only now
@Novitsky, Yes, if I first gunzip it, (and it renames it foo.tar) and then tar -tf foo.tar (unzip it again) it works. I didn't know I didn't need to rename it. Thanks for that tip.
I just also submitted a bug report to Firefox. But make no mistake, both your server and Firefox have issues, just different ones. If the file had not been double zipped in the first place I don't think Firefox would have gotten it wrong. But given that Chrome and wget get it right even when double zipped, I suggested to Firefox to change their guess and not jump the the conclusion that they don't need to unzip it.
Thanks again.
bryan or adam.donnison can you take a look at this? It seems Firefox is having trouble with how our webserver is serving gzipped files.
dbart I've turned off the gzip compression at the web server level and tested on FF and it now appears to work OK. Seems there may be a bug in that module as it was not supposed to encode anything other than text, JSON or XML.
Turns out it is in the method that the backend supplies files to the web server, they are being mis-typed and therefore the gzip handling assumes they require gzipping. As mentioned I've disabled gzip just on those files, and will work with bryan to put in a more permanent fix.
This should now be resolved. bryan has modified the backend code, and the web server is now correctly identifying the file as already compressed and is no longer double-compressing. Tested in FF and Chrome.
Great! Works here. Thanks. And interesting to hear what it turned out to be, and how it was fixed.
adam.donnison & bryan: Thank you for getting it fixed!