Pro ject : Main­tain your Mac’s stora ge

Beat per­for­mance is­sues to the punch

Mac Format - - CONTENTS -

Stor­age just stores, but to keep re­trieval re­li­able it needs to be

cared for. Over the long term, files pro­gres­sively get messier and re­quire pe­ri­odic tidy­ing. While manag­ing what’s stored is an im­por­tant first step in main­tain­ing stor­age, that isn’t our fo­cus here. In­stead, we’ll con­cen­trate on manag­ing how your files are stored and main­tain­ing the stor­age medium it­self, whether it’s a hard disk, solid-state drive (SSD), or a com­bi­na­tion. You can, of course, keep most of your files and their back­ups in re­mote cloud stor­age, but the cost of leav­ing oth­ers to do your main­te­nance is high, and you’re re­liant on its avail­abil­ity.

Our over­rid­ing con­cerns are per­for­mance and re­li­a­bil­ity. Left un­tended, hard disks can be­come pro­gres­sively slower to ac­cess, and im­pact on a Mac’s per­for­mance in use. This is be­com­ing in­creas­ingly con­tro­ver­sial: at one time many users be­lieved that if they did not de­frag­ment (de­frag) the files on their drives ev­ery month or so, ev­ery­thing would grind to a halt. Im­prove­ments in disk sys­tems and the chang­ing de­mands of OS X now make this a ques­tion­able prac­tice, pro­vided that suf­fi­cient free space is kept to al­low the ready cre­ation and use of cache and other tem­po­rary files.

OS X also per­forms its own back­ground de­frag­men­ta­tion when pos­si­ble, and old ways of forc­ing it to hap­pen have be­come more dif­fi­cult. If you try to clone your startup disk to an ex­ter­nal drive and back again – for­merly an ef­fec­tive way to con­sol­i­date free space – the re­sult may not start up at all. Only if your startup vol­ume is a hard disk, and you’re cer­tain its per­for­mance has fallen due to frag­men­ta­tion on it, should you con­sider try­ing to de­frag­ment it your­self.

Deal with dif­fer­ent medi­ums

Ul­ti­mately, all stor­age sys­tems de­velop spo­radic er­rors and may fail cat­a­stroph­i­cally. Plan­ning for those events is the ba­sis for keep­ing back­ups, and de­ter­mines the stor­age medium used for them. If you could hap­pily see all your stored doc­u­ments and files blown away at an in­stant, you do not need a backup,

As stor­age ac­cu­mu­lates

waste, its per­for­mance can drop off

but most of us need good and ex­ten­sive fall­backs, for which Time Ma­chine run­ning to an ex­ter­nal drive is usu­ally ideal.

As far as main­te­nance is con­cerned, there are now four main types of lo­cal stor­age: hard disks; SSDs; Fu­sion Drives, which in­cor­po­rate both a hard disk and an SSD work­ing to­gether in a Core Stor­age Log­i­cal Vol­ume Group; and net­work-at­tached stor­age (NAS).

A hard disk spins plat­ters coated with mag­netic ma­te­ri­als inside a sealed en­clo­sure, and is con­se­quently li­able to me­chan­i­cal and elec­tri­cal fail­ure, and to data cor­rup­tion. It un­der­goes max­i­mum stress when be­ing spun up from rest, so tends to last longer when left spin­ning con­stantly. It is also sen­si­tive to high tem­per­a­tures, and a disk run­ning hot is more likely to fail.

Though more usu­ally used as sin­gle units, mul­ti­ple hard disks can also be in­stalled in larger en­clo­sures to be used in­di­vid­u­ally (JBOD) or in con­cert as a RAID ar­ray. RAID used to be a spe­cial­ist pur­suit, but is in­creas­ingly com­mon. It is not an in­stant so­lu­tion, as there are dif­fer­ent lev­els with per­for­mance and ro­bust­ness trade-offs. Many peo­ple find RAID 5 across four hard disks is an ideal com­pro­mise, as it al­lows one of the disks to fail with­out any loss of data. Sim­ple mir­ror­ing, RAID 1, pro­vides re­dun­dancy but may im­pair per­for­mance, es­pe­cially when im­ple­mented in soft­ware. Hard­ware RAID im­ple­men­ta­tions are more ro­bust and faster than soft­ware RAID, but also more costly.

Hard disks usu­ally fail once their war­ranty pe­riod has ex­pired, com­monly af­ter four or more years. Those from the same batch tend to fail around the same time, so it’s worth sourc­ing from dif­fer­ent batches when build­ing an ar­ray, or two could fail in a short pe­riod. Hard disks older than about five years can last longer, but are worth re­plac­ing be­fore fail­ure.

SMART mon­i­tor­ing (Self-Mon­i­tor­ing, Anal­y­sis, and Re­port­ing Tech­nol­ogy) aims to de­tect early warn­ing signs of fail­ure, and you should al­ways take steps to re­place a hard disk with a wor­ry­ing sta­tus be­fore com­plete fail­ure oc­curs. How­ever, OS X only sup­ports sta­tus

DriveDx gives health and per­for­mance rat­ings for your drives, an SSD life­time in­di­ca­tor, and can email you about prob­lems.

Fu­sion Drives per­form well for their low cost, but they’re usu­ally hard to re­pair when they get into trou­ble.

Newspapers in English

Newspapers from Australia

© PressReader. All rights reserved.