1 |
On Sun, Jul 31, 2011 at 10:02 AM, Volker Armin Hemmann |
2 |
<volkerarmin@××××××××××.com> wrote: |
3 |
> On Friday 29 July 2011 14:18:41 Michael Mol wrote: |
4 |
>> Something that's been tickling my brain for a couple years now, and |
5 |
>> you guys are probably the right ones to ask. |
6 |
>> |
7 |
>> I haven't dropped coin for an SSD (yet), but I was wondering about |
8 |
>> uses for them beyond using them for / or /home. |
9 |
>> |
10 |
>> 1) What about sitting swap (partition, file, whatever) on the SSD? |
11 |
> |
12 |
> NO! |
13 |
> |
14 |
> For $DEITY's sake- NO! |
15 |
> |
16 |
> ssds can't withstand many writes (yeah, I know, millions blablabla... earlier |
17 |
> done than you think). Do Not Do This. |
18 |
> |
19 |
> SSDs are not meant for such a scenario. |
20 |
|
21 |
I'm not one to care what a tool was meant for, only what it can be used for. |
22 |
|
23 |
While I take your point about write-cycle limitations, and I would |
24 |
*assume* you're familiar with the various improvements on |
25 |
wear-leveling technique that have happened over the past *ten years* |
26 |
since those concerns were first raised, I could probably raise an |
27 |
argument that a fresh SSD is likely to last longer as a swap device |
28 |
than as a filesystem. |
29 |
|
30 |
Swap is only touched as-needed, while there's been an explosion in |
31 |
programs and user software which demands synchronous writes to disk |
32 |
for data integrity purposes. (Firefox uses sqlite in such a way, for |
33 |
example; I discovered this when I was using sqlite heavily in my *own* |
34 |
application, and Firefox hung for a couple minutes during every batch |
35 |
insert.) |
36 |
|
37 |
Also, despite the MBTF data provided by the manufacturers, there's |
38 |
more empirical evidence that the drives expire faster than expected, |
39 |
anyway. I'm aware of this, and not particularly concerned about it. |
40 |
|
41 |
> |
42 |
>> Presumably, in scenarios where expanding the RAM in a system is |
43 |
>> prohibitively expensive, an SSD could reduce the impact of swap |
44 |
>> thrash. |
45 |
> |
46 |
> no, it is increasing the impact of SSD trash. |
47 |
|
48 |
False dichotomy. Yes, it increases the wear on the device. That says |
49 |
nothing of its impact on system performance, which was the nature of |
50 |
my point. |
51 |
|
52 |
> |
53 |
>> 2) While my system rarely goes above using 2-2.5GB of RAM, I enjoy |
54 |
>> having 6-8GB of RAM, just for the file cache. Of course, I lose that |
55 |
>> when I reboot; the cache needs to be repopulated. Has there been any |
56 |
>> work in the kernel for doing things like Vista/Win7's ReadyBoost? |
57 |
>> ReadyBoost has a ridiculous limit to only using 4GB of a flash drive, |
58 |
>> but I'd think that an 80GB SSD would be a massive performance |
59 |
>> improvement. |
60 |
>> |
61 |
> |
62 |
> with a SSD filecache is not that important anymore - and every usb-stick is |
63 |
> slower than a SSD. |
64 |
|
65 |
I'll poke the second argument first. I wouldn't use USB for something |
66 |
like this. USB2 is a painfully slow polling architecture. Something |
67 |
like this would need to be done with SATA. I'm *not* that daft. |
68 |
|
69 |
As for a filecache not being that important, that's only the case if |
70 |
your data of interest exists on the filesystem you put on the SSD. |
71 |
|
72 |
Let's say you're someone like me, who would tend to go with 60GB for / |
73 |
and 3TB for /home. At various times, I'll be doing HDR photo |
74 |
processing, some video transcoding, some random non-portage compile |
75 |
jobs, web browsing, coding, etc. |
76 |
|
77 |
If I take a 160GB SSD, I could put / (or, at least, /var/ and /usr), |
78 |
and have some space left over for scratch--but it's going to be a pain |
79 |
trying to figure out which of my 3TB of /home data I want in that fast |
80 |
scratch. |
81 |
|
82 |
File cache is great, because it takes caches your most-used data from |
83 |
*anywhere* and keeps it in a fast-access datastore. I could have a 3 |
84 |
*petabyte* volume, not be particularly concerned about data |
85 |
distribution, and have just as response from the filecache as if I had |
86 |
a mere 30GB volume. Putting a filesystem on an SSD simply cannot scale |
87 |
that way. |
88 |
|
89 |
Actually, this conversation reminds me of another idea I'd had at one |
90 |
point...putting ext3/ext4's journal on an SSD, while keeping the bulk |
91 |
of the data on large, dense spinning platters. |
92 |
|
93 |
> |
94 |
>> Obviously, for something like Gentoo, putting an SSD-based filesystem |
95 |
>> under /var/tmp makes a lot of sense, but what other uses have been |
96 |
>> tried? How'd they work out? |
97 |
> |
98 |
> no, /var/tmp is very not important from a performance point of view - with the |
99 |
> exception of /var/tmp/portage - and that is a candidate for tempfs. |
100 |
|
101 |
Did you miss the last week's worth of discussion of memory limits on tmpfs? |
102 |
|
103 |
-- |
104 |
:wq |