Intel® Quartus® Prime Software
Intel® Quartus® Prime Design Software, Design Entry, Synthesis, Simulation, Verification, Timing Analysis, System Design (Platform Designer, formerly Qsys)
16612 Discussions

Request: Generate zsync error recovery files for any/all Quartus 11 release files

Altera_Forum
Honored Contributor II
1,310 Views

Hello, 

 

I'm trying to download Quartus 11, and have encountered a md5sum mismatch (relative to the listed value on the Altera web site) for the downloaded file despite using a download manager and having apparently 'successfully' downloaded the file of the expected size and name. This has happened occasionally particularly with large multi-gigabyte files which are extremely slow and unreliable to transfer with my particular internet connection situation. It is frustrating to have to start downloading over from zero having a downloaded file which is almost certainly 99.99999% correct but apparently with small pieces of corruption of unknown sizes/locations within the received files. Obviously I'd like to improve my ISP connection and isolate the cause of this corruption (it is possibly a fault in the remote server or download manager software, or possibly some ISP or OS/SW related problem), but for the moment I'd just like an easy way to recover from detected errors in downloaded files that may take several real time days to attempt to re-download (with a not so good probability of success at that...). 

 

It is common practice in many large software/data distributions to generate some kind of ECC or block level HASH / checksum files to accompany large files such that any missing/corrupt pieces can efficiently be identified and corrected with only a small incremental download of the bad blocks. 'par' parity files, zsync, jigdo, rsync, torrent files, and other such commonly used systems enable this efficient error detection / correction functionality. 

 

Particularly, it would be a trivial thing to use the 'zsync' program (commonly available for UNIX/LINUX) to generate a small (few megabytes) hash file that contains sufficient block level error detection data to facilitate recovery of corrupted transfers. This auxiliary file is calculated operating on any chosen known good original large file (e.g. a quartus distribution file or disc image). 

 

Here's a working shell command line example how one might generate a "zsync" hash file based on a local copy of a particular distribution file: 

zsyncmake -v -e -C 11.0_acds_web_edition.iso 

That operates on the master file named 11.0_acds_web_edition.iso, and generates the file: 

11.0_acds_web_edition.iso.zsync which is a few megabytes in size, and contains the relevant 

error detection data allowing for incremental downloading of any problematic parts given that the 

.zsync file generated is shared with anyone that may have a need to do such corrections. 

Typically a software vendor would gebnerate the zsync files based on each of the large released files, and make these .zsync files also available for download in the software distribution, though anyone having 

a good copy of a released file could independently generate and share this hash data for the benefit of 

others. 

 

Ideally Altera's release process could eventually be modified to automatically generate and distribute these or similar kinds of files; sharing the md5sum as is currently done is great to enable people to see if they have a corruption problem, but if they do, there's no efficient way to FIX it without such additional data as are here proposed. 

 

Anyway I've seen some comments about people getting Quartus 11 et. al. working under UBUNTU or similar UNIX OSs, so perhaps someone who has some of these files already successfully downloaded and a UNIX workstation might consider making a zsync file out of whatever version (they're all 2+ GBy) of source files you happen to have downloaded, and share the zsync file(s) via something like an "shared material" alteraforum attachment or via some file hosting service (windows live skydrive, mediafire, rapidshare, whatever? 

 

Or if someone on the ALTERA LAN wants to add this as a convenience to your users, it should only take a few minutes to generate them with something like: 

mkdir zsync_files && cd zsync_files && find /path_to_master_release_files/ -type f -print0 | xargs -0 -n1 zsyncmake -v -e -C 

cp -p *.zsync /path/to/shared/files/on/public/server/ 

...or otherwise upload/share the .zsync files. 

 

(and, yes, I know, one may request the DVD, though in this case it seems that it'll take around a month to be available, hence I thought I'd ask to see if someone wanted to make the correction files available sooner...) 

 

q.v. url for more info on zsync: 

zsync.moria.org.uk

 

 

Thanks in advance!
0 Kudos
5 Replies
Altera_Forum
Honored Contributor II
617 Views

i built the file for you. if you could put up an AlteraWiki page explaining what you said here and including the link, you might encourage people to fill in the rest later 

 

http://alterawiki.com/wiki/file:11.0_acds_web_edition.iso.zsync.tgz
0 Kudos
Altera_Forum
Honored Contributor II
617 Views

 

--- Quote Start ---  

i built the file for you. if you could put up an AlteraWiki page explaining what you said here and including the link, you might encourage people to fill in the rest later 

 

http://alterawiki.com/wiki/file:11.0_acds_web_edition.iso.zsync.tgz 

--- Quote End ---  

 

 

Thank you very much for generating that zsync file! 

 

Using it, in just a few moments zsync assured me of the surprising result that the file I had already downloaded was 100% identical to the file you used to generate the zsync file, so it was quite effective in resolving any doubts that I had encountered errors / corruption in the transfer / storage of the download. 

The problem, it seems, is that the file is either corrupt on Altera's server itself, or that Altera has posted the incorrect MD5 check value for this particular file since our two copies of the file have identical SHA1s, but my local MD5 for the file doesn't match the advertised value from Altera's web site. 

 

Anyway you've saved me a day or more of fruitless re-downloading, many thanks! 

 

Official MD5 for file listed at Altera WWW download site: 

www.altera.com/download/software/quartus-ii-we 

ACDS v11.0 Web Edition DVD (ISO) Windows/Linux 11.0_acds_web_edition.iso 

MD5: 45dc537fbf1bd99eb76564c53d79f2c5 

 

 

Downloaded copy local MD5 (and SHA1 for what it is worth) results: 

md5sum --binary 11.0_acds_web_edition.iso 

65b93c4da718557447b9ca5c90057727 *11.0_acds_web_edition.iso 

 

sha1sum 11.0_acds_web_edition.iso 

461c546275173edfc088e640b2d72a947f35f809 11.0_acds_web_edition.iso
0 Kudos
Altera_Forum
Honored Contributor II
617 Views

my MD5 matches yours. davka explained the issue in the other thread 

 

65b93c4da718557447b9ca5c90057727 11.0_acds_web_edition.iso 

 

still, an interesting idea with FEC for huge downloads
0 Kudos
Altera_Forum
Honored Contributor II
617 Views

Thanks again, all's well that ends well. I'll try to take your suggestion and share the suggestion and any further zsync file resources I can orchestrate via the wiki for the benefit of others / future use cases.

0 Kudos
Altera_Forum
Honored Contributor II
617 Views

par2 seems like a similar tool but has better cross platform support: 

 

http://www.quickpar.org.uk/
0 Kudos
Reply