1 /* $FreeBSD: src/sys/dev/isp/Hardware.txt,v 1.2 2003/01/01 18:48:50 schweikh Exp $ */
3 Hardware that is Known To or Should Work with This Driver
8 This is not an endorsement for hardware vendors (there will be
9 no "where to buy" URLs here with a couple of exception). This
10 is simply a list of things I know work, or should work, plus
11 maybe a couple of notes as to what you should do to make it
12 work. Corrections accepted. Even better would be to send me
13 hardware to I can test it.
15 I'll put a rough range of costs in US$ that I know about. No doubt
16 it'll differ from your expectations.
23 There are various suffices that indicate copper or optical
24 connectors, or 33 vs. 66MHz PCI bus operation. None of these
25 have a software impact.
27 Approx cost: 1K$ for a 2200
31 These are the new 2-Gigabit cards. Optical only.
36 Antares P-0033, P-0034, P-0036
38 There many other vendors that use the Qlogic 2X00 chipset. Some older
39 2100 boards (not on this list) have a bug in the ROM that causes a
40 failure to download newer firmware that is larger than 0x7fff words.
42 Approx cost: 850$ for a P-0036
46 In general, the 2200 class chip is to be preferred.
53 Of the two, the 1000 (7 ports, vs. 12 ports) has had fewer problems-
54 it's an old workhorse.
57 Approx cost: 1.5K$ for Vixel 1000, 2.5K$ for 2000
59 Gadzoox Cappellix 3000
60 Don't forget to use telnet to configure the Cappellix ports
61 to the role you're using them for- otherwise things don't
64 (cost: I have no idea... certainly less than a switch)
70 (other brocades should be fine)
72 Especially with revision 2 or higher f/w, this is now best
73 of breed for fabrics or segmented loop (which Brocade
76 For the Silkworm II, set operating mode to "Tachyon" (mode 3).
78 The web interace isn't good- but telnet is what I prefer anyhow.
80 You can't connect a Silkworm II and the other Brocades together
81 as E-ports to make a large fabric (at least with the f/w *I*
82 had for the Silkworm II).
84 Approx cost of a Brocade 2400 with no GBICs is about 8K$ when
85 I recently checked the US Government SEWP price list- no doubt
86 it'll be a bit more for others. I'd assume around 10K$.
90 This also is a fine switch, but you have to use a browser
91 with working java to manage it- which is a bit of a pain.
92 This also supports fabric and segmented loop.
94 These switches don't form E-ports with each other for a larger
99 McData (model unknown)
101 I tried one exactly once for 30 minutes. Seemed to work once
102 I added the "register FC4 types" command to the driver.
104 (cost: very very expensive, 40K$ plus)
108 Multimode optical is adequate for Fibre Channel- the same cable is
109 used for Gigabit Ethernet.
111 Copper DB-9 and Copper HSS-DC connectors are also fine. Copper &&
112 Optical both are rated to 1.026Gbit- copper is naturally shorter
113 (the longest I've used is a 15meter cable but it's supposed to go
116 The reason to use copper instead of optical is that if step on one of
117 the really fat DB-9 cables you can get, it'll survive. Optical usually
118 dies quickly if you step on it.
120 Approx cost: I don't know what optical is- you can expect to pay maybe
121 a 100$ for a 3m copper cable.
125 I use Finisar copper and IBM Opticals.
127 Approx Cost: Copper GBICs are 70$ each. Opticals are twice that or more.
130 Vendor: (this is the one exception I'll make because it turns out to be
131 an incredible pain to find FC copper cabling and GBICs- the source I
132 use for GBICs and copper cables is http://www.scsi-cables.com)
136 There now is apparently a source for little connector boards
137 to connect to bare drives: http://www.cinonic.com.
140 5. Storage JBODs/RAID
144 Rinky-tink, but a solid 4 bay loop only entry model.
146 I paid 1000$ for mine- overprice, IMO.
150 I rather like this box. The blue LEDs are a very nice touch- you
151 can see them very clearly from 50 feet away.
153 I paid 2000$ for one used.
157 Very expensive (in my opinion) but well crafted. Has two SES
158 instances, so you can use the ses driver (and the example
159 code in /usr/share/examples) for power/thermal/slot monitoring.
161 Approx Cost: The last I saw for a price list item on this was 22K$
162 for an unpopulated (no disk drive) A5X00.
165 DataDirect E1000 RAID
167 Don't connect both SCSI and FC interfaces at the same time- a SCSI
168 reset will cause the DataDirect to think you want to use the SCSI
169 interface and a LIP on the FC interface will cause it to think you
170 want to use the FC interface. Use only one connector at a time so
171 both you and the DataDirect are sure about what you want.
173 Cost: I have no idea.
177 This is a software storage virtualization engine that
178 runs on Sparc/Solaris in target mode for frontend
179 and with other FC or SCSI as the backend storage. FreeBSD
180 has been used extensively to test it.
183 Cost: I have no idea.
187 I have used lots of different Seagate and a few IBM drives and
188 typically have had few problems with them. These are the bare
189 drives with 40-pin SCA connectors in back. They go into the JBODs
192 Seagate does make, but I can no longer find, a little paddleboard
193 single drive connector that goes from DB-9 FC to the 40-pin SCA
194 connector- primarily for you to try and evaluate a single FC drive.
196 All FC-AL disk drives are dual ported (i.e., have separate 'A' and
197 'B' ports- which are completely separate loops). This seems to work
198 reasonably enough, but I haven't tested it much. It really depends
199 on the JBOD you put them to carry this dual port to the outside
200 world. The JMR boxes have it. The Sun A5X00 you have to pay for
201 an extra IB card to carry it out.
203 Approx Cost: You'll find that FC drives are the same cost if not
204 slightly cheaper than the equivalent Ultra3 SCSI drives.
206 7. Recommended Configurations
208 These are recommendations that are biased toward the cautious side. They
209 do not represent formal engineering commitments- just suggestions as to
210 what I would expect to work.
212 A. The simpletst form of a connection topology I can suggest for
213 a small SAN (i.e., replacement for SCSI JBOD/RAID):
216 2xxx <----------> Single Unit of Storage (JBOD, RAID)
218 This is called a PL_DA (Private Loop, Direct Attach) topology.
220 B. The next most simple form of a connection topology I can suggest for
221 a medium local SAN (where you do not plan to do dynamic insertion
222 and removal of devices while I/Os are active):
225 2xxx <----------> +--------
235 This is a Private Loop topology. Remember that this can get very unstable
236 if you make it too long. A good practice is to try it in a staged fashion.
238 It is possible with some units to "daisy chain", e.g.:
241 2xxx <----------> (JBOD, RAID) <--------> (JBOD, RAID)
243 In practice I have had poor results with these configurations. They *should*
244 work fine, but for both the JMR and the Sun A5X00 I tend to get LIP storms
245 and so the second unit just isn't seen and the loop isn't stable.
247 Now, this could simply be my lack of clean, newer, h/w (or, in general,
248 a lack of h/w), but I would recommend the use of a hub if you want to
249 stay with Private Loop and have more than one FC target.
251 You should also note this can begin to be the basis for a shared SAN
252 solution. For example, the above configuration can be extended to be:
255 2xxx <----------> +--------
264 2xxx <----------> +--------
266 However, note that there is nothing to mediate locking of devices, and
267 it is also conceivable that the reboot of one host can, by causing
268 a LIP storm, cause problems with the I/Os from the other host.
269 (in other words, this topology hasn't really been made safe yet for
272 D. You can repeat the topology in #B with a switch that is set to be
273 in segmented loop mode. This avoids LIPs propagating where you don't
274 want them to- and this makes for a much more reliable, if more expensive,
277 E. The next level of complexity is a Switched Fabric. The following topology
278 is good when you start to begin to get to want more performance. Private
279 and Public Arbitrated Loop, while 100MB/s, is a shared medium. Direct
280 connections to a switch can run full-duplex at full speed.
283 2xxx <----------> +---------
292 2xxx <----------> +---------
295 I would call this the best configuration available now. It can expand
296 substantially if you cascade switches.
298 There is a hard limit of about 253 devices for each Qlogic HBA- and the
299 fabric login policy is simplistic (log them in as you find them). If
300 somebody actually runs into a configuration that's larger, let me know
301 and I'll work on some tools that would allow you some policy choices
302 as to which would be interesting devices to actually connect to.