Recovery within long-running transactions

Abstract

As computer systems continue to grow in complexity, the possibilities of failure increase. At the same time, the increase in computer system pervasiveness in day-to-day activities brought along increased expectations on their reliability. This has led to the need for effective and automatic error recovery techniques to resolve failures. Transactions enable the handling of failure propagation over concurrent systems due to dependencies, restoring the system to the point before the failure occurred. However, in various settings, especially when interacting with the real world, reversal is not possible. The notion of compensations has been long advocated as a way of addressing this issue, through the specification of activities which can be executed to undo partial transactions. Still, there is no accepted standard theory; the literature offers a plethora of distinct formalisms and approaches.In this survey, we review the compensations from a theoretical point of view by: (i) giving a historic account of the evolution of compensating transactions; (ii) delineating and describing a number of design options involved; (iii) presenting a number of formalisms found in the literature, exposing similarities and differences; (iv) comparing formal notions of compensation correctness; (v) giving insights regarding the application of compensations in practice; and (vi) discussing current and future research trends in the area.

Recovery for Failures in Rolling Upgrade on Clouds

“…One difference is that DDG is automatically generated during the process execution [19], while in our research the expected system global states are defined according to operation requirements. DDG is serving as the main system contextual knowledge which is used by process recovery actions such as undo and redo [17] [25], and our expected system global states are the main contextual knowledge that is used by the two recovery strategies proposed by us for the recovery of cloud rolling upgrade operation.…”

“…For long running transactions, recovery strategies usually involve backward recovery and forward recovery [17]. Backward recovery refers to the strategy which first reverts the current erroneous state to a previous correct state before attempting to continue execution.…”

“…Our recovery strategy of Compensated Undo & Redo takes similarity to backward recovery and our recovery strategy of Reparation takes similarity to forward recovery. Another form of forward recovery in long running transactions is called compensation [17], which means to attempt to correct the state of a system given some knowledge of the previous actions of the system [17]. Generally, our recovery strategies take similarity to the recovery mechanisms for long running transactions but there are some challenges introduced such as state reachability check.…”

Towards a Taxonomy of Cloud Recovery Strategies

“…long running transactions [17], usually involve backward recovery and forward recovery [17]. Backward recovery refers to the strategy which first revert the current erroneous state to a previous correct state before attempting to continue execution.…”

“…One form of backward recovery is called rewind & replay [26], which makes the system go back to the previous consistent state and replay the step in the transaction. One form of forward recovery in long running transactions is called compensation [17], which means to attempt to correct the state of a system given some knowledge of the previous actions of the system [17]. When conducting sporadic operations such as rolling upgrade for a large-scale cloud application, backward recovery and forward recovery can be employed for recovering from errors happening during the operations.…”