Coerced Cache Eviction and discreet mode journaling: Dealing with misbehaving disks

Rajimwale, Abhishek; Chidambaram, Vijay; Ramamurthi, Deepak; Arpaci-Dusseau, Andrea C.; Arpaci-Dusseau, Remzi H.

doi:10.1109/dsn.2011.5958264

Cited by 7 publications

(4 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The durability of written data can be improved if one alternatively disables the on-disk cache, applies flush workloads to the cache, or uses controllers with batterybacked cache [Nightingale et al 2006;Rajimwale et al 2011;Shin et al 2011]. In most of our experiments we kept enabled the on-disk caches, but in Section 5.6 we report the sensitivity of our results to alternative cache configurations.…”

Section: Performance Evaluationmentioning

confidence: 88%

“…For that purpose, the Linux ext3 provides the barrier mount option while the SCSI specification offers the SYNCHRONIZE CACHE command [SBC 2005] (SATA also provides the FLUSH CACHE command [SATA 2003]). If the device does not support write barriers, a flush workload can be used to flush the on-disk write cache instead [Rajimwale et al 2011]. Alternatively, we can disable the write cache and have the disk only acknowledge a write after it really reaches the medium [Shin et al 2011].…”

Section: Atomicitymentioning

confidence: 99%

“…If it involves a large amount of written data or numerous small writes, double writes lead to inefficient utilization of storage bandwidth that may hurt performance. Consequently, the use of data journaling is explicitly discouraged in several cases, while production systems activate only metadata journaling by default [Anand et al 2008;Chidambaram et al 2012;Prabhakaran et al 2005a;Rajimwale et al 2011;Zhang and Ghose 2007].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Improving Bandwidth Efficiency for Consistent Multistream Storage

Hatzieleftheriou

Anastasiadis

2013

ACM Trans. Storage

View full text Add to dashboard Cite

Synchronous small writes play a critical role in system availability because they safely log recent state modifications for fast recovery from crashes. Demanding systems typically dedicate separate devices to logging for adequate performance during normal operation and redundancy during state reconstruction. However, storage stacks enforce page-sized granularity in data transfers from memory to disk. Thus, they consume excessive storage bandwidth to handle small writes, which hurts performance. The problem becomes worse, as filesystems often handle multiple concurrent streams, which effectively generate random I/O traffic. In a journaled filesystem, we introduce wasteless journaling as a mount mode that coalesces synchronous concurrent small writes of data into full page-sized journal blocks. Additionally, we propose selective journaling to automatically activate wasteless journaling on data writes with size below a fixed threshold. We implemented a functional prototype of our design over a widely-used filesystem. Our modes are compared against existing methods using microbenchmarks and application-level workloads on stand-alone servers and a multitier networked system. We examine synchronous and asynchronous writes. Coalescing small data updates to the journal sequentially preserves filesystem consistency while it reduces consumed bandwidth up to several factors, decreases recovery time up to 22%, and lowers write latency up to orders of magnitude.

show abstract

Section: Performance Evaluationmentioning

confidence: 88%

Section: Atomicitymentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Improving Bandwidth Efficiency for Consistent Multistream Storage

Hatzieleftheriou

Anastasiadis

2013

ACM Trans. Storage

View full text Add to dashboard Cite

show abstract

“…Примером может служить отличие в работе системного вызова fsync в Linux и в Mac OS X. В случае последней вызова fsync недостаточно для обеспечения долговечности, требуется установка дополнительного параметра через вызов fcntl (F_FULLFSYNC) [3]. Более того, некоторые жесткие диски игнорируют команду переноса данных из кеша в основной памяти (flush) на постоянный носитель, чтобы показывать себя лучше в тестах производительности [4]. Еще один пример -отличие семантики системного вызова fdatasync в Linux [5] и FreeBSD [6].…”

Section: Introductionunclassified

Design and Implementation of a Tool for Testing Stored Data Durability for Applications Based on File Systems

Rodionov¹,

Кузнецов²

2023

Proceedings of ISP RAS

View full text Add to dashboard Cite

Applications that work with data are required to ensure their reliable storage. The interfaces available for working with file systems are not sufficiently specified and require high qualifications for correct use that does not lead to loss of user data. As part of this work, a tool was developed that provides developers with the opportunity to test their applications and identify the most common errors. The tool is based on collecting events from the interaction of the application with the file system and then running checks that can indicate errors. The tool implements a modular architecture that allows you to expand the available set of checks. The developed tool was integrated into the process of testing the implementation of a durable log, similar to the write ahead log, a component implemented in many database management systems. The tool allowed to detect and correct several errors leading to possible data loss.

show abstract

Reliability Analysis of SSDs Under Power Fault

Zheng

Tucek

Qin

et al. 2016

ACM Trans. Comput. Syst.

View full text Add to dashboard Cite

Modern storage technology (solid-state disks (SSDs), NoSQL databases, commoditized RAID hardware, etc.) brings new reliability challenges to the already-complicated storage stack. Among other things, the behavior of these new components during power faults—which happen relatively frequently in data centers—is an important yet mostly ignored issue in this dependability-critical area. Understanding how new storage components behave under power fault is the first step towards designing new robust storage systems. In this article, we propose a new methodology to expose reliability issues in block devices under power faults. Our framework includes specially designed hardware to inject power faults directly to devices, workloads to stress storage components, and techniques to detect various types of failures. Applying our testing framework, we test 17 commodity SSDs from six different vendors using more than three thousand fault injection cycles in total. Our experimental results reveal that 14 of the 17 tested SSD devices exhibit surprising failure behaviors under power faults, including bit corruption, shorn writes, unserializable writes, metadata corruption, and total device failure.

show abstract

Coerced Cache Eviction and discreet mode journaling: Dealing with misbehaving disks

Cited by 7 publications

References 21 publications

Improving Bandwidth Efficiency for Consistent Multistream Storage

Improving Bandwidth Efficiency for Consistent Multistream Storage

Design and Implementation of a Tool for Testing Stored Data Durability for Applications Based on File Systems

Reliability Analysis of SSDs Under Power Fault

Contact Info

Product

Resources

About