| 1 |
On 08/09/2016 00:12, Volker Armin Hemmann wrote: |
| 2 |
> Am 07.09.2016 um 08:18 schrieb Alan McKinnon: |
| 3 |
>> On 07/09/2016 01:57, Volker Armin Hemmann wrote: |
| 4 |
>>> Am 01.09.2016 um 11:01 schrieb Alan McKinnon: |
| 5 |
>>>> On 01/09/2016 09:18, gevisz wrote: |
| 6 |
>>>>> 2016-09-01 9:13 GMT+03:00 Alan McKinnon <alan.mckinnon@×××××.com>: |
| 7 |
>>>>>> On 01/09/2016 08:04, gevisz wrote: |
| 8 |
>>>> [snip] |
| 9 |
>>>> |
| 10 |
>>>>>> it will take about 5 seconds to partition it. |
| 11 |
>>>>>> And a few more to mkfs it. |
| 12 |
>>>>> Just to partition - may be, but I very much doubt |
| 13 |
>>>>> that it will take seconds to create a full-fledged |
| 14 |
>>>>> ext4 file system on these 5TB via USB2 connention. |
| 15 |
>>>> Do it. Tell me how long it tool. |
| 16 |
>>>> |
| 17 |
>>>> Discussing it without doing it and offering someone else's opinion is a |
| 18 |
>>>> 100% worthless activity |
| 19 |
>>>> |
| 20 |
>>>>> Even more: my aquiantance from the Window world |
| 21 |
>>>>> that recomended me this disc scared me that it may |
| 22 |
>>>>> take days... |
| 23 |
>>>> Mickey Mouse told me it takes microseconds. So what? |
| 24 |
>>>> |
| 25 |
>>>> Do it. Tell me how long it took. |
| 26 |
>>>> |
| 27 |
>>>>>>> Is it still advisable to partition a big hard drive |
| 28 |
>>>>>>> into smaller logical ones and why? |
| 29 |
>>>>>> The only reason to partition a drive is to get 2 or more |
| 30 |
>>>>>> smaller ones that differ somehow (size, inode ratio, mount options, etc) |
| 31 |
>>>>>> |
| 32 |
>>>>>> Go with no partition table by all means, but if you one day find you |
| 33 |
>>>>>> need one, you will have to copy all your data off, repartition, and copy |
| 34 |
>>>>>> your data back. If you are certain that will not happen (eg you will |
| 35 |
>>>>>> rather buy a second drive) then by all means dispense with partitions. |
| 36 |
>>>>>> |
| 37 |
>>>>>> They are after all nothing more than a Microsoft invention from the 80s |
| 38 |
>>>>>> so people could install UCSD Pascal next to MS-DOS |
| 39 |
>>>>> I definitely will not need more than one mount point for this hard drive |
| 40 |
>>>>> but I do remember some arguments that partitioning a large hard drive |
| 41 |
>>>>> into smaller logical ones gives me more safety in case a file system |
| 42 |
>>>>> suddenly will get corrupted because in this case I will loose my data |
| 43 |
>>>>> only on one of the logical partitions and not on the whole drive. |
| 44 |
>>>>> |
| 45 |
>>>>> Is this argument still valid nowadays? |
| 46 |
>>>> That is the most stupid dumbass argument I've heard in weeks. |
| 47 |
>>>> It doesn't even deserve a response. |
| 48 |
>>>> |
| 49 |
>>>> Who the fuck is promoting this shit? |
| 50 |
>>>> |
| 51 |
>>>> |
| 52 |
>>> people who had to deal with corrupted filesystems in the past? |
| 53 |
>>> |
| 54 |
>>> |
| 55 |
>> The way to deal with the problem of fs corruption is to have reliable |
| 56 |
>> tested backups. |
| 57 |
>> |
| 58 |
>> The wrong way to deal with the problem of fs corruption is to get into |
| 59 |
>> cargo-cult manoeuvrers thinking that lots of little bits making a whole |
| 60 |
>> is going to solve the problem. |
| 61 |
>> |
| 62 |
>> Especially when the part of the disk statistically most at risk is the |
| 63 |
>> valuable data itself. OS code can be rebuilt easily, without backups |
| 64 |
>> data can't. |
| 65 |
>> |
| 66 |
> |
| 67 |
> the bigger the drive, the greater the chance of fs corruption. Just by |
| 68 |
> statistics. Better one minor partition is lost than everything. |
| 69 |
|
| 70 |
What are the statistical chances of that one minor partition being the |
| 71 |
one that gets corrupted? Statistically the odds are very small. |
| 72 |
|
| 73 |
Think about it, if the minor partition is say 5% of the disk and if all |
| 74 |
other things are exactly equal, the odds are 1 in 20. |
| 75 |
|
| 76 |
Apart from inherent defects in the drive itself, the sectors that are |
| 77 |
more prone to failing are those that are read the most and to a larger |
| 78 |
extent those that are written the most. |
| 79 |
|
| 80 |
What is read the most? OS and Data |
| 81 |
What is written the most? Data |
| 82 |
What has by far the greatest likelihood of suffering fs corruption? Data |
| 83 |
|
| 84 |
|
| 85 |
|
| 86 |
> |
| 87 |
> You can disagree as much as you like, but with the size of drives and |
| 88 |
> the current error rate of consumer hard drives it is not a question of |
| 89 |
> 'if' but just a matter of 'when'. |
| 90 |
> |
| 91 |
|
| 92 |
I don't disagree with you. I'm disagreeing with cargo cult mentality |
| 93 |
that dividing a disk up into lots of smaller partitions somehow |
| 94 |
magically confers significant safety margins of some magical kind. Go |
| 95 |
read the OPs opening statement again, he's quoting a friend from 20 |
| 96 |
years ago and the statement consists entirely of woo-woo magic |
| 97 |
hand-wavey statements, the kind of shit I have to deal with every day |
| 98 |
from twits with just enough IQ to read executive white papers. |
| 99 |
|
| 100 |
Yes, drives fail. Yes, consumer drives are crap. With 3TB now being |
| 101 |
common place and prices plunging, we have 20G or so for OS and 2980GB |
| 102 |
full of data. That 20G is so small and immaterial in terms of risk we |
| 103 |
can just disregard it and assume the only thing that can be damaged is |
| 104 |
2980G of data. |
| 105 |
|
| 106 |
Solution: back up the whole damn lot properly and forget what we did 20 |
| 107 |
years ago. That was farting in a breeze, nowadays it's farting in a |
| 108 |
hurricane. |
Archives