Automating Live Update for Generic Server Programs

Abstract

The pressing demand to deploy software updates without stopping running programs has fostered much research on live update systems in the past decades. Prior solutions, however, either make strong assumptions on the nature of the update or require extensive and error-prone manual effort, factors which discourage the adoption of live update. This paper presents Mutable Checkpoint-Restart (MCR), a new live update solution for generic (multiprocess and multithreaded) server programs written in C. Compared to prior solutions, MCR can support arbitrary software updates and automate most of the common live update operations. The key idea is to allow the running version to safely reach a quiescent state and then allow the new version to restart as similarly to a fresh program initialization as possible, relying on existing code paths to automatically restore the old program threads and reinitialize a relevant portion of the program data structures. To transfer the remaining data structures, MCR relies on a combination of precise and conservative garbage collection techniques to trace all the global pointers and apply the required state transformations on the fly. Experimental results on popular server programs (Apache httpd, nginx, OpenSSH and vsftpd) confirm that our techniques can effectively automate problems previously deemed difficult at the cost of negligible performance overhead (2% on average) and moderate memory overhead (3.9x on average, without optimizations).

Formal methods in dynamic software updating: a survey

“…In some papers, the process of formalisation is done after the development of the system (Noubissi et al, 2011;Chen et al, 2007;Zhang et al, 2012) whereas in other papers, this formal study is done before or during the process of programming (Murarka, 2010;Hayden et al, 2012;Anderson, 2013;Stoyle, 2006). In this trend, research papers pointed recently the importance to build update aware applications (Giuffrida and Tanenbaum, 2009;Giuffrida et al, 2017), the authors used the expression live update-friendly systems to describe this idea. In An et al (2015), Zhang et al (2014Zhang et al ( , 2015 and Wu et al (2008), the authors used formal methods at design level.…”

Evolving NoSQL Databases without Downtime

“…Our work is also related to the body of research on dynamic software updates [20], [23], [21], [22], which aim to modify a running program on-the-fly, without causing downtime. However, with the exception of a position paper [55], these approaches focus on changes to code and data structures loaded in memory, rather than changes to the formats of persistent data stored in a database.…”

“…When a data upgrade is installed, applications with an incompatible version must update themselves. They can do this either by simple stop-and-restart (to the new application version), or they can use dynamic software updating (DSU) [20], [21], [22], [23] or concurrent application switching (as in parallel AppEngine [24]) to avoid lost application state and/or shorten pause times. KVolve triggers conversions automatically as data is accessed.…”

Dynamic and coordinated software reconfiguration in distributed data stream systems

“…Many stream processing systems have to provide services for 24x7, with no acceptable downtime [1,2]. However, they commonly have to cope with changes in their execution environment (e.g., moving from onpremises architecture to cloud architecture or changing the network technology) and in the requirements that they must comply with [3] (e.g., adding new functionality or modifying existing parts).…”

“…Despite extensive research in dynamic software reconfiguration [8][9][10][11], safe reconfiguration is still an open problem [2,[12][13][14]. A common approach is to put the component that has to be updated into a safe state, such as the quiescent state [15], before reconfiguring the system [16].…”