Skip to content

ImAWolf/ncbi-vdb

 
 

Repository files navigation

The NCBI SRA (Sequence Read Archive)

Contact:

email: sra-tools@ncbi.nlm.nih.gov or visit the SRA Tools web site

Release 2.4.4

This release contains improvements in the area of networking. A bug in our networking code was introduced recently that allowed HTTP server connection resets to go undetected, which manifested itself as a "corrupt blob" error. What was happening was that the download-on-demand code was properly detecting a dropped connection, but failing to report it to consuming code. The latter in turn saw blobs of invalid data, and correctly reported that they were corrupt.

The most common manifestation of this problem was with fastq-dump, which seemed to encounter a corrupt blob error after dumping some number of reads. The reported problem was correct. However, the source of the error was not a corrupt download, but instead was the result of dropped connections with the HTTP server.

This release corrects the problem with unreported cases of dropped connections, and so restores the expected behavior.

NCBI-VDB

The NCBI-VDB library implements a highly compressed columnar data warehousing engine that is most often used to store genetic information. Databases are stored in a portable image within the file system, and can be accessed/downloaded on demand across HTTP.

For additional information on using, configuring, and building the library, please visit our wiki: https://github.com/ncbi/ncbi-vdb/wiki.

SRA Toolkit Development Team

Packages

No packages published

Languages

  • C 74.5%
  • Assembly 16.0%
  • C++ 6.5%
  • Shell 0.9%
  • Makefile 0.8%
  • Perl 0.5%
  • Other 0.8%