1 |
Branko Badrljica <brankob@××××××××××.com> posted |
2 |
49DE848D.60905@××××××××××.com, excerpted below, on Fri, 10 Apr 2009 |
3 |
01:28:13 +0200: |
4 |
|
5 |
> Duncan wrote: |
6 |
> |
7 |
> |
8 |
>> Did you try md-mod,start_dirty_degraded=1 (AFAIK this applies to |
9 |
>> RAID-4/5/6 only)? |
10 |
>> |
11 |
> No |
12 |
|
13 |
Hmm, and you said RAID-5, right? |
14 |
|
15 |
I think one of the reasons I may have never had to use that option on |
16 |
RAID-6 is that it has two-way parity, so chances are either one or the |
17 |
other will still be valid -- even if it crashes in the middle of a write, |
18 |
one stripe-write might be invalid, but as long as I wasn't already |
19 |
running degraded, the state on the others should be clean, either clean- |
20 |
state from before the write, or clean-state from after the write, thus |
21 |
ensuring data integrity. Naturally, due to hardware cache buffers, etc, |
22 |
this isn't entirely guaranteed (unless the write buffers are disabled), |
23 |
but the chances are far higher of a clean state than they are with RAID-5 |
24 |
with its single-way parity. |
25 |
|
26 |
That's one reason that at least after I groked RAID-4/5/6 a bit better, |
27 |
I've been keen to get back a fully functional RAID-6 after a crash with |
28 |
one of the disks returning unclean on reboot, thus degrading the RAID |
29 |
until it is recovered. In I believe two cases I've degraded to *NO* |
30 |
redundancy, dropping TWO disks out due to not catching the one dropped |
31 |
out for some time (this was before I really groked this RAID stuff, so I |
32 |
wasn't being quite as careful). Had I crashed again before recovery of |
33 |
at least one of the drives in those cases, thus before getting back at |
34 |
least single-way-parity, chances are I'd have had to use the |
35 |
start_dirty_degraded option, with its corresponding risk of lost data |
36 |
(the reason for the option and its defaulting OFF) due to failure of the |
37 |
data integrity guarantee RAID-4/5/6 normally includes. |
38 |
|
39 |
(FWIW I do run without UPS. Unfortunately the reality is that a proper |
40 |
UPS for this level of machine, at least back when I was running CRTs, |
41 |
would cost near what the machine itself did, and given that I was already |
42 |
buying near the top of my budget range that put a UPS out of |
43 |
consideration. Since I switched to LCD I've been considering it again |
44 |
and would likely have one by now had the economy and therefore my budget |
45 |
not tanked.) |
46 |
|
47 |
At least as I understand it, given RAID-5 only, loss of one spindle as |
48 |
"dirty" is fine, but if the system crashes while recovering it and the |
49 |
write-intent bits indicate an uncompleted write on the two remaining |
50 |
spindles, that option is needed as the alternative would be loss of the |
51 |
data integrity guarantee. With a RAID-6, there's the additional |
52 |
protection of two-way parity and thus a second spindle can drop out |
53 |
without killing the data integrity guarantee, but the risk remains the |
54 |
same if there's a crash in that state and a /third/ drops out as unclean. |
55 |
|
56 |
So I guess I'm glad I decided to go with the additional protection of |
57 |
RAID-6, even if I did so before I fully groked what I was doing. |
58 |
|
59 |
>> What about listing the appropriate component devices, as so: |
60 |
>> md=d1,/dev/sda1,/dev/sdb1,/dev/sdc1... ? |
61 |
>> |
62 |
> Yes. I had it by default. Without it never seemed to work. With it, it |
63 |
> worked _sometimes_. |
64 |
|
65 |
Wow. I guess under those conditions I'd probably be going initramfs |
66 |
too. Fortunately my experience has been far more robust, even without a |
67 |
UPS. If that has been due to my choice of RAID-6, over a RAID-5 with a |
68 |
hot-spare, I guess I'm glad I made the cautious choice even if I didn't |
69 |
fully groke the implications back when I made it. |
70 |
|
71 |
> Initramfs seem to be superior solution. With it there is more stuff I |
72 |
> can do, like trying to rebuild array, fsck filesystems thoroughly etc. |
73 |
> |
74 |
> In fact, this is what it does now. Amongst other things, it checks |
75 |
> /etc/fstab, gets filesystem types for "/boot" and "/" and e2fscks them, |
76 |
> if needed. |
77 |
|
78 |
That's what my "rootbak" partition is for. =:^) |
79 |
|
80 |
> Being early adopter of ext4 I was many times in situation when boot-time |
81 |
> e2fsck found some serious errors but demanded to be rerun on unmounted |
82 |
> media to be able to finish. Without initramfs, only way to do this is |
83 |
> through CD, but kernel there doesn't recognise disks on my Dell. Besides |
84 |
> that it is tedious. |
85 |
|
86 |
Well, you /could/ burn a CD with a kernel built with the proper options, |
87 |
but you are correct, booting from a CD is indeed a hassle. |
88 |
|
89 |
But again, a "rootbak" copy of your root partition to fall back on would |
90 |
work just as well as the initramfs, and backups are something people |
91 |
should be doing anyway, so for me anyway, it was less (as in no |
92 |
additional) work. |
93 |
|
94 |
> That thingie is _definitely_ very useful and I plan to make it |
95 |
> permanent. |
96 |
|
97 |
Indeed, it sounds like it. Of course, a fully functional second copy of |
98 |
the root partition would be even more useful. =:^) |
99 |
|
100 |
> I had completed the script that does the packing in maybe two hours ( I |
101 |
> don't remeber anymore ) and it needs maybe a few minutes to make initrd, |
102 |
> copy it to /boot/initrd and old one to /boot/initrd_old. Also, if I need |
103 |
> one more package in the initrd ( say new version of gcc or python etc), |
104 |
> it is only matter of adding package name in one line... |
105 |
> |
106 |
> But I will do something much more compact in C. With maybe an option to |
107 |
> do "DefCon 4" style boot from USB key toolshed. |
108 |
> |
109 |
> I found USB keys to be useful for that role, when opened and with |
110 |
> soldered "in-line" conenctor. With it it becomes small dongle that is |
111 |
> plugged directly into motherboard and so practically part of the |
112 |
> machine. |
113 |
> |
114 |
> Tha way, you get "Terminator style" sturdy machine that is capable of |
115 |
> reboot no matter what just as motherboard, CPU and PSU stay functional. |
116 |
> ;o) |
117 |
|
118 |
Indeed. This sounds actually very practical. I believe there's actually |
119 |
motherboards that do it, but it's a high-end feature, with a roll-your- |
120 |
own being much more flexible and much cheaper, for those willing to do it. |
121 |
|
122 |
Actually, I've seen mods for *-WRT class routers and software (OpenWRT, |
123 |
DD-WRT, Tomato, etc) that do pretty much the same thing, solder a USB |
124 |
based flash module to the thing and expand their storage and flexibility |
125 |
by an order or two of magnitude! Run a server, etc, directly off your |
126 |
router! They can be made detachable, or not, as desired. |
127 |
|
128 |
> Yeah, but what do you do if your raid becomes non-functional for some |
129 |
> reason ? |
130 |
> |
131 |
> I have four disks and two RAIDs. First is boot bartition RAID-1, which |
132 |
> ebncompasess firs 1GB of each disk, second is RAID-5, which has the rest |
133 |
> of the capacity. |
134 |
> |
135 |
> All partitions are typed as "fd"- self-detect RAID. On boot it doesn't |
136 |
> matter. BIOS and GRUB see first partitions as if they were ordinary |
137 |
> partitions with ext2 that happen to have exactly the same content. |
138 |
> |
139 |
> So, if one or more of the disks failed, I could boot with initrd from |
140 |
> whatever disk is left... |
141 |
> |
142 |
> While in initramfs, I check and assemble both arrays and exec |
143 |
> "make-init" ( or whatever its name it is- that util in klibc- it is |
144 |
> effectively thingie that erases everything in /rootfs, remounts new root |
145 |
> to / and starts init ) |
146 |
|
147 |
Here, it's three RAIDs on 4 disks/spindles, but all three covering the |
148 |
same four spindles, with one each RAID-0/1/6, each running in parallel |
149 |
partitions on each of the four spindles. |
150 |
|
151 |
I had initially intended to put recovery software on the RAID-1, and |
152 |
sized it accordingly. However, after I was setup and it was time to do |
153 |
so I decided I didn't actually need it. |
154 |
|
155 |
But if I were doing it over, as I said earlier, I'd definitely create |
156 |
separate RAID-6s instead of the single partitioned RAID-6, and use two |
157 |
rootbaks so three root filesystem images total, one on each of three |
158 |
separate RAID-6s, so if one failed to come up, I'd just change the |
159 |
command line as necessary to bring up the second, and if it failed, the |
160 |
third. |
161 |
|
162 |
I'd also split the RAID-1, either creating two single-disk RAID-1s or |
163 |
more likely, four separate /boot images direct on the individual spindles. |
164 |
|
165 |
I don't think I'd do initramfs, but something that just occurred to me |
166 |
now -- I could actually copy the content of an entire rescue CD to one of |
167 |
the four disks (leaving three as /boot copies), setting it up as an |
168 |
alternative boot I could access directly from BIOS, to recover the rest |
169 |
of the system, should I need to. |
170 |
|
171 |
>> I won't be doing nVidia any time soon. Not until they come around to |
172 |
>> properly cooperating with the FLOSS community. YMMV and I know a lot |
173 |
>> of gamers especially value their games above their freedom. That's not |
174 |
>> my life and not my systems and therefore not my call, but what I run |
175 |
>> here IS, and if I wanted to be stuck with proprietaryware, I'd have not |
176 |
>> bothered dumping a decade's worth of experience on MS to start over |
177 |
>> with Linux! |
178 |
>> |
179 |
> Well, at least every fresh version of their driver is available on |
180 |
> Gentoo at the same instant it gets out. Even without it, nvidia's own |
181 |
> installer can figure things out. |
182 |
> |
183 |
> Not so with Radeon. I needed 9.1 version of driver when latest in gentoo |
184 |
> tree was 8.1. I downloaded the thing and died wrestling with damn |
185 |
> scripts halfway through. |
186 |
|
187 |
The community wisdom has been this: |
188 |
|
189 |
1) If you want an expansion card and freedomware drivers, go with (older) |
190 |
Radeons. For many years it was r2xx based chips at the latest so 92xx |
191 |
series cards. Now the r3xx-r5xx series, pretty much all the xXXX (x300- |
192 |
x1900) cards, are quite well supported, including 3D. The r600 and r700 |
193 |
series, basically the hd* cards, aren't as well supported /just/ yet, but |
194 |
development is coming on strong since the specs were released and they |
195 |
look to be reasonably well supported by this time next year. |
196 |
|
197 |
2) For the gamers out there and anyone else needing the features only to- |
198 |
date available in proprietaryware only drivers due to lack of specs (or |
199 |
with the later Radeons, only recently available specs), nVidia's the way |
200 |
to go. I believe it's also the choice for integrated video |
201 |
proprietaryware folks, tho I'm not as familiar with that market and |
202 |
honestly don't know. ATI's proprietary drivers have been improving |
203 |
recently (since AMD took them over and they turned over a new leaf in |
204 |
regard to Linux), but for years were, plainly stated, crap. While nVidia |
205 |
has refused to open the specs, they HAVE been reasonably good -- by far |
206 |
and away the best -- with their closed source Linux drivers, so there's |
207 |
no question, that's the way to go if you don't care about the |
208 |
proprietaryware aspect and are willing to jump thru the hoops and accept |
209 |
the delays it forces on users and the community in general. |
210 |
|
211 |
3) For integrated use with freedomware drivers, in the historical mid- |
212 |
term, Intel was the top choice and remains a good one. Of course, that |
213 |
means Intel systems as well, AMD fans need not apply, but that's the way |
214 |
it is. Since AMD bought ATI and with the resulting changes there, in |
215 |
theory they should have a working integrated video AMD/ATI platform |
216 |
freedomware solution either new or more likely coming, but I really don't |
217 |
know. In the laptop/integrated market today, I'd personally chose Intel, |
218 |
no question about it. |
219 |
|
220 |
> I don't get what kind of machine did they write that thing for- it sure |
221 |
> as hell wasn't anything like mine. |
222 |
> Also, when it comes to OpenGL performance, nVidia is still the king as I |
223 |
> understand. |
224 |
|
225 |
Yes. As I said, ATI's proprietaryware Linux drivers have had the |
226 |
reputation as crap. If you are doing proprietaryware, until /very/ |
227 |
recently (as in this year), nVidia, hands-down. That said, ATI's latest |
228 |
have improved a lot, from what I've read, to the point Phoronix among |
229 |
others seems to be recommending them evenly with nVidia now. /That/ |
230 |
said, I'm honestly not interested enough in the proprietaryware side to |
231 |
be the one to talk to about this. But if you're in the market for brand |
232 |
new higher end graphics cards right now, do at least take a look at the |
233 |
available Phoronix.com resources. |
234 |
|
235 |
> I have upgraded my dual setup to three-monitor setup ( 3x Samsung 204B, |
236 |
> 3x1600x1200) and have had to add elcheapo 9400 card to existing dual-DVI |
237 |
> 8800GT. |
238 |
> |
239 |
> Do you know about AMD/ATI Linux drivers- do they do such kind of stuff |
240 |
> without extra wizardry ? |
241 |
|
242 |
I don't know much about the proprietaryware side at all. At present, I |
243 |
don't know that much about the newer hardware freedomware drivers either, |
244 |
tho I expect I'll know much more later this year as I'm just now starting |
245 |
to look at upgrading my aging Radeon 9200, since there's actually decent |
246 |
support for newer stuff now. |
247 |
|
248 |
What I've seen so far is as I said above, thru the r500 series, basically |
249 |
the Radeon xXXXX cards, support is now close to that on the older cards |
250 |
and since they're newer, that means better/more-capabilities in many |
251 |
cases. But for the current r600 and r700 parts, basically hd* cards, |
252 |
support is under fast and furious development but is lagging six months |
253 |
minimum, more like nine. Still, by next spring support should be quite |
254 |
good. |
255 |
|
256 |
I expect I'll be buying a near top of the series r500 xXXXX card, here, |
257 |
for three reasons. (1) Support is more solid currently. (2) I'm still |
258 |
running AGP and PCI-X, NOT PCI-E (FWIW, this is the last major system |
259 |
upgade I plan on this thing, tho I expect to be running it several years |
260 |
still), so for me, the benefits (or even availability) of the r600/r700 |
261 |
chip HD* cards aren't likely to be that great over the cheaper r500 chip |
262 |
xXXXX cards. (3) I could never justify $400+ for a graphics card in any |
263 |
case, and as a couple generations out, r5xx chip xXXXX cards should be |
264 |
cheap in comparison to the r6xx/r7xx chip HD* cards. |
265 |
|
266 |
Specifically to your question, I really don't know on the multi-card |
267 |
thing. I had it working years ago (xf86 era, with the nVidia card and |
268 |
proprietary driver and an old 4 MB S3 Virge), but have stuck with single- |
269 |
card dual-out solutions since upgrading to dual 21/22" CRTs @ 1600x1200 |
270 |
each some years ago, and later to dual 1920x1200 LCDs, currently. (I |
271 |
wanted the 30" 2560x1600 LCDs, but $1000+ each plus a new card at the |
272 |
same time, to handle dual-dual-link-DVI, wasn't in the cards. There's |
273 |
always the next upgrade cycle, I guess.) I /had/ tried it with early |
274 |
xorg era and Radeon r1xx/r2xx based cards and freedomware drivers, but |
275 |
didn't get it working. I've not tried it since, however, as resolutions |
276 |
and screen sizes have improved to the point where I've not found it as |
277 |
pressing as it once was. |
278 |
|
279 |
I read somewhere that support for multi-card had broken at one point in |
280 |
the transfer from the old multi-screen entry layout section |
281 |
hard-xorg.conf-configured setups to the newer dynamic PnP xrandr based |
282 |
setups, but I don't know if that was xorg-wide, just for that driver, |
283 |
just for xrandr at that point, or what. I also have little idea whether |
284 |
it has been fixed, altho obviously if you're running three monitors on |
285 |
two cards with an nVidia driver, it still works with that. |
286 |
|
287 |
Anyway, ask me in Sept. or Nov. or so, and I'll probably be far more |
288 |
informed, as I'll be fresh off of doing all the research necessary to |
289 |
determine what I want for an upgrade, instead of just starting it. =:^) |
290 |
|
291 |
-- |
292 |
Duncan - List replies preferred. No HTML msgs. |
293 |
"Every nonfree program has a lord, a master -- |
294 |
and if you use the program, he is your master." Richard Stallman |