Talk:Disk mirroring/Archive 1

Merge with Disk mirror
I can't see why these should be two separate articles, so have recommended them for merging. Thoughts? --JennyRad 18:13, 26 June 2006 (UTC)

Merge with Disk mirroring
Can't see why these should be two separate articles so have recommended a merge. Thoughts? --JennyRad 18:14, 26 June 2006 (UTC)
 * Agreed --Pboyd04 02:27, 20 July 2006 (UTC)
 * Actually on second thought... merge them both to RAID. --Pboyd04 02:28, 20 July 2006 (UTC)

Mirroring is the same as replication?
The terminology in storage world goes IMHO like this:


 * mirroring - local RAID 1


 * synchronous replication - similar to mirroring, but:
 * usually longer distance
 * write operation is guaranteed to be completed on both sides or not at all (!)


 * asynchronous replication - this is very rarely called mirroring


 * BCV aka flashsnap - this is never called mirroring, because it employs non-trivial techniques of quick resync after split. And the term "third mirror" implies full resync after split - much cheaper, simpler and more time consuming.

This article uses much broader definition of mirroring and mixes very different concepts. Therefore I think page should be renamed to Disk Storage Replication, with trivial mirroring stuff moved to RAID and eventually BCV stuff to a new page. --Kubanczyk 11:29, 17 March 2007 (UTC)

Make one General and other sub-articles
Yes! It could be merged but in other case you can put a general article on just Mirroring and other sub-articles joined with that general article about different types of mirroring used in computer world. — Preceding unsigned comment added by 193.251.135.125 (talk) 18:29, 8 May 2007 (UTC)

Logical Difference between Mirroring & Replication
Replication means just another copy of data placed on an other computer. They didn't have any relation except that they have identical files(At the time of just being created). But mirroring means that if a data stored on Computer A has it's replications on Computer B, C and D then all four A, B, C and D are connected with each other through network and any change in data at Computer A will also made effect on Computer B, C and D. So, in mirroring It's true that you make replication of data BUT it's not simply a replication. After replication they keeps in touch with each other and try their best to serve identical data by updating with original source when changes are made to data on the main computer. At the end in two line Replication means just creating a copy of data on other computer and Mirroring means not only creating a copy of data but also update them just after making some changes to data to keep both sides identical as long as they exists. —The preceding unsigned comment was added by 193.251.135.125 (talk) 18:42, 8 May 2007 (UTC).


 * Stop right there. This is your definition of replication. I think opinion of industry leaders might be a little more useful (EMC, HDS, IBM, Sun, Veritas, NetApp, ...). Since storage terminology is still pretty fuzzy whoever wins the market sets the global standard. The current winner seems to agree with me, since in SRDF the letter R is for Replication :) NetApp also agrees. Veritas too. --Kubanczyk 10:52, 9 May 2007 (UTC)

Only 2 disks?
The article implies that only two disks may be mirrored, with maybe one spare for a total of three drives. I think this is inaccurate. RAID1 allows many disks to be created inside a mirroring configuration (factually these use concatenation + mirroring), so why wouldn't other disk mirror schemes allow this? FluffleS 00:54, 13 June 2007 (UTC)

defines RAID 1 as having exactly 2 disks. Assuming this to be an authorative statement (I've not found anything in the website to be false to me, but I also have no conclusive proof to take it so), mirroring plus concatenation would be more a case of several RAID 1 arrays put into a JBOD array for the concatenation. 67.170.192.5 00:23, 24 June 2007 (UTC)


 * I would not consider such a web site as authoritative. Although typically a basic mirrored solution uses one pair of disks, there is no requirement. RAID is built on mirroring data multiple times for error correction, so sticking to the article being around the basic concepts, the article needs to allow for multiple disks.

Grammar
IanMSpencer undid the changes that used "Data" a a plural noun. Rather than undo his undo, I just comment here. A subjective statement such as '"data are" does not read well' doesn't stand up well against objective facts of grammar. In fact, to many, "data is" is just plain grating and distracting. In any case, most technical journals require that "data" be used as a plural noun (see the Wikipedia on Data), and this is a technical article. Rknasc 21:19, 10 September 2007 (UTC)

Data is an uncounted noun. Correct grammar is that an uncounted noun is treated as singular (especially in US English, there is a little more leeway in British English but still correct and common usage). In other words it is not a subjective judgement but sound grammar (and English grammar rarely manages to be objective or factual - try explaining as to why something happens on Christmas Day, at Christmas or in Winter - we don't do logic!). Spenny 22:49, 10 September 2007 (UTC)

PS, try this for a ref which covers some of the well known facts about such things. Spenny 22:53, 10 September 2007 (UTC)

...and as I am in an anecdotal mood, my first letter published in a national newspaper, the Independent, was in response to some old cove who claimed that learning Latin was the solution to all the modern spelling problems. I responded "As I recall, all this meant to me and my peers was that we couldn't spell in two languages." (I was inordinately pleased with myself for that one). Spenny 22:59, 10 September 2007 (UTC)

From the Wikipedia page on Data "Many (perhaps most) academic, scientific, and professional style guides (e.g., see page 43 of the World Health Organization Style Guide) request that authors treat data as a plural noun." Whatever. Rknasc 13:27, 11 September 2007 (UTC)


 * Wikipedia is not a citable source. Also if you read that page you will realise that it is not consistent and the page itself cannot help but wander from one form to another as different editors have had a go. Try quoting the opening sentence: when the data becomes organized it becomes information. - do you really believe that is improper English? My take is that the contention is that the scientific world has hung onto the plural usage, but the main world does not. In fact, I have put a note that the Data page needs to be updated to properly reflect there are two usages, both of which are acceptable, though there is a leaning to plurality in the scientific world. In the computing world, it is the more common usage that dominates and it is simply not reflecting the reality. Dictionaries should describe usage not dictate it. The fact that style guides have this as a commonly confused situation is perhaps a hint that there is a disagreement.
 * Traditionally and in technical use data is treated as a plural, as in Latin it is the plural of datum. In modern non-scientific use, however, it is often treated as a singular, and sentences such as data was collected over a number of years are now acceptable. - Compact Oxford Dictionary. As a graduate in Computer Science, a published author, and 25 years of experience, I can assure you that it is a rarity to hear anyone in IT, either here in the UK or in the USA talk of data as anything but as a singular form (and I have no problem with computing not being a science!). Spenny 21:18, 11 September 2007 (UTC)

Difference between Mirroring, Failover, Clustering and Load balancing?
What is the difference between Disk mirroring, Failover,  High-availability clusters (a.k.a. Clustering?)  and  Load balancing?

I would like to see a concise comparison of concepts across all of these pages, as all of these concepts seems tightly bound, and perhaps some of them are identical.

--Eptin 21:59, 9 September 2007 (UTC)

Replication_(computer_science) on Wikipedia confirms that 500m is far too low a limit for 'fast enough for mirroring' - with the use of Fiber Channel over WDM you should be able manage 10's or 100's of kilometres. The only limitation is how much of a slowdown for disk writes is acceptable - smart use of data journalling for filesystems and DBMSs should be able to cope with higher disk latencies since fewer disk writes are required to write a transaction. —Preceding unsigned comment added by 81.2.106.242 (talk) 18:20, 6 October 2007 (UTC)

"Pair" or "Disk"
The article includes the clause, "The copying of data from one pair of a mirror to another..." Should this say "disk" instead of "pair"? — Preceding unsigned comment added by 148.87.67.207 (talk) 19:56, 22 March 2012 (UTC)

Web mirror redirects here?
Web mirror #REDIRECTs here, but disk mirroring and web mirrors are quite different things. This article (the disk mirroring article) does not mention Web mirrors. What should be done about this? --Jeff Ogden (W163) (talk) 18:20, 28 January 2014 (UTC)


 * Nice catch! Just  this redirect page so it now points to Load balancing (computing), what's pretty much it.  Hope you agree. &mdash; Dsimic (talk | contribs) 23:57, 28 January 2014 (UTC)
 * I do agree. Thanks. --Jeff Ogden (W163) (talk) 05:45, 29 January 2014 (UTC)