1 |
Hi guys, after quick read about ssd, I have a couple of question: |
2 |
1. My friend have new server with a ssd installed. He plan to RHEL 5.7 |
3 |
(I don't know why he choose this) on it. On redhat website, it say |
4 |
something like this: |
5 |
"However, if the device does not export topology information, Red Hat |
6 |
recommends that the first partition be created at a 1MB boundary." |
7 |
What does it mean by 1MB boundary? Does it mean he have to create 1MB |
8 |
free space in front or he have to create a 1MB partition in front of |
9 |
his actual partition(s)? |
10 |
|
11 |
2. Is it possible to combine TRIM support and ext3 partition (AFAIK, |
12 |
RHEL 5.7 haven't support ext4)? |
13 |
|
14 |
*i hope this is not count as hijacking |
15 |
|
16 |
On 8/14/12, Alan McKinnon <alan.mckinnon@×××××.com> wrote: |
17 |
> On Mon, 13 Aug 2012 11:55:31 -0400 |
18 |
> Michael Mol <mikemol@×××××.com> wrote: |
19 |
> |
20 |
>> On Mon, Aug 13, 2012 at 11:47 AM, Alan McKinnon |
21 |
>> <alan.mckinnon@×××××.com>wrote: |
22 |
>> |
23 |
>> > On Mon, 13 Aug 2012 08:17:23 -0400 |
24 |
>> > Michael Mol <mikemol@×××××.com> wrote: |
25 |
>> > |
26 |
>> > > On Mon, Aug 13, 2012 at 4:06 AM, Neil Bothwick |
27 |
>> > > <neil@××××××××××.uk> wrote: |
28 |
>> > > |
29 |
>> > > > On Sun, 12 Aug 2012 14:11:37 -0400, Allan Gottlieb wrote: |
30 |
>> > > > |
31 |
>> > > > > > I have one of those. But I decided to stick with |
32 |
>> > > > > > traditional DOS partitioning style and grub instead of GPT |
33 |
>> > > > > > and grub2. |
34 |
>> > > > > |
35 |
>> > > > > I am leaning toward traditional partitioning, but with |
36 |
>> > > > > grub2. Do those two not mix well? |
37 |
>> > > > |
38 |
>> > > > GRUB2 works fine with MBR partition tables. But if you're |
39 |
>> > > > starting from scratch, you may as well use GPT and get rid of |
40 |
>> > > > the legacy MBR limitations and fragility. |
41 |
>> > > > |
42 |
>> > > |
43 |
>> > > I'm not dissing GPT...but what's fragile about MBR? |
44 |
>> > |
45 |
>> > it's 30 years old, |
46 |
>> > only 4 primary partitions, |
47 |
>> > only 16 extended partitions, |
48 |
>> > it's got that weird DOS boot flag thing, |
49 |
>> > it all has to fit in one sector. |
50 |
>> > |
51 |
>> > I had to fix a mispartitioned disk over the weekend, this really |
52 |
>> > should have been a simple mv-type operation, but because all 4 |
53 |
>> > primary partitions were in use I had to disable swap and use it as |
54 |
>> > a leap-frog area. It felt like I was playing 15 pieces with the |
55 |
>> > disk. That's fragile - not that the disk breaks, but that it breaks |
56 |
>> > my ability to set the thing up easily. |
57 |
>> > |
58 |
>> > Basically, mbr was built to cater for the needs of DOS-3. In the |
59 |
>> > meantime, 1982 called and they want their last 30 years back. |
60 |
>> > |
61 |
>> > Just because we can hack workarounds into it to get it to function |
62 |
>> > doesn't mean we should continue to use it. |
63 |
>> > |
64 |
>> |
65 |
>> You misunderstand me. I wasn't arguing that GPT wasn't perhaps more |
66 |
>> elegant than MBR and dos partitions. I wanted to know what was |
67 |
>> _fragile_ about MBR. Completely different things. |
68 |
> |
69 |
> I did answer (somewhat obliquely). |
70 |
> |
71 |
> mbr as a single isolated unit is not especially fragile; very little |
72 |
> software is and bits don't magically "rot" |
73 |
> |
74 |
> It's the system into which the sysadmin inserts mbr that is fragile. |
75 |
> The whole system is fragile like an egg is fragile - it can't withstand |
76 |
> much manhandling or moving of stuff around before some mistake wreaks |
77 |
> everything, and that is mostly due to mbr's limits. |
78 |
> |
79 |
> It's not semantic nitpicking here, if the system as a unit becomes |
80 |
> fragile as a result of part X, then the system is still fragile. |
81 |
> |
82 |
> -- |
83 |
> Alan McKinnon |
84 |
> alan.mckinnon@×××××.com |
85 |
> |
86 |
> |
87 |
> |
88 |
|
89 |
|
90 |
-- |
91 |
Salam, |
92 |
|
93 |
J.Marcos S. |
94 |
Sent from X1™ |