Handling Duplicates in Dockerfiles Families: Learning from Experts

Abstract

Docker is becoming a popular tool used by developers and end-users to deploy and run software applications. Dockerfiles are now found alongside projects' source code. Several projects are even starting to maintain families of Dockerfiles, like the Python project that maintains a family of 43 Dockerfiles, each for a specific Python version on a specific Linux distribution. In some situations, Dockerfiles family maintainers have to propagate a change to several, if not all, Dockerfiles of the family (for instance a bugfix applying on all Dockerfiles targeting Python 2.7). This need to propagate changes is usually due to the presence of duplicates between several Dockerfiles of the family. In this paper, our goal is to provide practitioners a clear explanation for why Dockerfile duplicates arise in projects, and what are the different means to handle duplicates with their pros and cons. To perform a grounded analysis, we observe the practices of expert Dockerfile maintainers of Official Docker projects, which are setting the best-practices for Dockerfile maintenance. We show that duplicates in Dockerfiles are frequent in our corpus, that maintainers are aware of their existence, are frequently facing them and have a mixed opinion regarding them (error-prone when not using any tool, but easy to maintain with the right tools). Finally, we describe and analyse the tools used by maintainers to handle duplicates.

An empirical study on self-admitted technical debt in Dockerfiles

“…While many studies focusing on Docker have been conducted (Haque et al 2020;Oumaziz et al 2019), as far as we know, there has been no study on SATD in the Docker domain. SATD may be one of the causes by which image quality deteriorates if developers do not repay the debt.…”

DRIVE: Dockerfile Rule Mining and Violation Detection

“…Similarly, shipwright [20] also attempts to repair the broken Dockerfiles to pass the building requirements through static analysis. Some other work has been proposed to address the duplicates or type-2 clone issues among multiple Dockerfiles [34,44]. Different from our work, the emphasis of such work is mainly to detect faults or duplicates instead of best practice violations.…”

Type-2 Code Clone Detection for Dockerfiles

“…There are a few studies on code clone detection in Dockerfiles. Oumaziz et al propose a duplicate code detection technique for Dockerfiles [10]. They use inverted index algorithm [19] with some Dockerfile instructions as a chunk to detect duplicate code after parsing the Dockerfiles.…”

“…This experiment targets 4,817 Dockerfiles contained in 725 repositories that are popular on GitHub. In Docker, there is a practice that the entire flow of container building is often described in a template file, within which the defining minor differences between distributions and versions in temporary variables that can be replaced with text are defined [10]. The target Dockerfiles are automatically generated by executing or compiling this template.…”

“…Clone detection techniques have been proposed for not only procedural languages such as Java and C but also various other languages. On the other hand, there have been a few proposals for Dockerfile clone detection In one such study, Oumaziz et al have proposed a duplicate code detection technique for Dockerfiles [10]. This technique isolates tokens in Dockerfiles and detects duplicates using inverted index algorithm.…”