or Connect
AVS › AVS Forum › Video Components › Home Theater Computers › Guide To Building A Media Storage Server
New Posts  All Forums:Forum Nav:

Guide To Building A Media Storage Server - Page 58

post #1711 of 7891
Quote:
Originally Posted by MikeSM View Post

It's an atheros gigabit ethernet chip (builtin to the P45 motherboard). As I said though, netperf works just fine...

thx
mike

Did you make the tcp global changes on Server 2008 as well?
post #1712 of 7891
Quick update...

I'm swapping emails w/ the LSI folks. Sent them log files from the 84016E and they think I may have a backplane and/or cable issue (those errors I had mentioned). I'm redoing the config at the moment and will post new benchmark results as warranted.

Also on a side note, don't know if anyone saw this, but Buy.com has added WD's 2TB drive to their site (not available at the moment tho):

http://www.buy.com/prod/3-5-2tb-sata...210673283.html

Sidenote #2: Also had another HD fail the other day -- a Seagate. Gotta say that Seagate's RMA/warranty program is MUCH better than the non-existent RMA process offered by Samsung (MWAVE is taking care of my bad Samsung drive) as UPS dropped off the replacement Seagate drive today and it also included a prepaid sticker for the return shipment. Nice.
post #1713 of 7891
More expensive then the 4020, you lose 4 bays, and your limited to micro-atx boards? It does look nice though. Looks like it can only support a micro-atx, doesn't say that in the specs though....hmm
post #1714 of 7891
2Tb drives are a bit pricey right now...hmm If they were ~200 it would be worthwhile as more capacity in the same space for around the same price as 1TB drives.
post #1715 of 7891
Quote:
Originally Posted by Dougie085 View Post

More expensive then the 4020, you lose 4 bays, and your limited to micro-atx boards? It does look nice though. Looks like it can only support a micro-atx, doesn't say that in the specs though....hmm

huh? It supports up to full e-atx motherboards.

I may have spoken too soon...Since it uses a standard ATX PSU (on it's side), the layout may not support ATX or higher motherboards. Hmm...That case does look nice. I may order one just to play with.
post #1716 of 7891
Mike, could you check on your 2008 server, in the physical nic settings if Large Send Offload v2 (IPv4) is enabled.

If you have Hyper-V installed, also check the virtual switch nic settings if
TCP Large Send Offload (IPv4) is enabled.



Quote:
Originally Posted by MikeSM View Post

Guys, I have a question for all your folks running Vista and Windows server 2008 together. I have my rebuilt raid server running server 2008 using an adaptec controller with 2 sets of 8 drives in raid6. Local performance is great, but I can't seem to break more than 35 MB/sec (over a switched GbE LAN) over CIFS to my primary desktop workstaion, which has 3 7200.10 500 GB drives in RAID0, and also has good local performance.

If I use netperf, I an saturate the GbE between the two machines (with a 64K window), so it's not a wiring or network issue.

Any ideas as to what could be going on? Vista has SP1 loaded, and multimedia network throttling turned off. All the tcp global params are set for speed. Receive scaling, chimney offload, ecn and rfc1323 enabled, and recieve autotuning set to normal and ctcp as addon congestion control provider.

What am I missing here?
post #1717 of 7891
Quote:
Originally Posted by plasticquart View Post

I'm swapping emails w/ the LSI folks. Sent them log files from the 84016E and they think I may have a backplane and/or cable issue (those errors I had mentioned). I'm redoing the config at the moment and will post new benchmark results as warranted.

Is that in relation to the PD xx errors you posted earlier?

I've been getting all sorts of errors (see below) on some Seagate 1.5TB drives connected through the Chenbro CK13601 and eventually the controller would give up and die on me requiring a reboot to come back to life. Pulled the 1.5TB drives from the expander and connected them straight to the controller and all is good. I'm gonna try to reflash the RAID card now and see if that makes a difference as well as try some smaller 500GB drives.

---errors---
01/22/09 15:40:48: Task Mgmt Start Addr a14d3400 DevId[1d] Index 0 chip 0 type 3 numCmdIssued 9
01/22/09 15:40:48: EVT#12829-01/22/09 15:40:48: 267=Command timeout on PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 2a 00 06 51 41 80 00 00 80 00
01/22/09 15:40:48: EVT#12830-01/22/09 15:40:48: 267=Command timeout on PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 2a 00 06 51 41 00 00 00 80 00
01/22/09 15:40:48: EVT#12831-01/22/09 15:40:48: 267=Command timeout on PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 2a 00 06 51 42 80 00 00 80 00
01/22/09 15:40:48: EVT#12832-01/22/09 15:40:48: 267=Command timeout on PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 2a 00 06 51 43 00 00 00 80 00
01/22/09 15:40:48: EVT#12833-01/22/09 15:40:48: 267=Command timeout on PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 2a 00 06 51 40 00 00 00 80 00
01/22/09 15:40:48: EVT#12834-01/22/09 15:40:48: 267=Command timeout on PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 2a 00 06 51 42 00 00 00 80 00
01/22/09 15:40:48: EVT#12835-01/22/09 15:40:48: 267=Command timeout on PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 2a 00 06 51 40 80 00 00 80 00
01/22/09 15:40:48: EVT#12836-01/22/09 15:40:48: 267=Command timeout on PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 2a 00 06 51 43 80 00 00 80 00
01/22/09 15:40:48: EVT#12837-01/22/09 15:40:48: 268=PD 1d(e0xff/s29) Path 5001c4500000f50c reset (Type 03)
01/22/09 15:40:48: MPI_EVENT_SAS_DISCOVERY: PortBitmap ff - Discovery is in progress
01/22/09 15:41:18: Disc-prog= 0....resetProg=0 aenCount=0 transit=0
01/22/09 15:41:19: EVT#12859-01/22/09 15:41:19: 113=Unexpected sense: PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 2a 00 06 65 f2 80 00 00 80 00, Sense: 70 00 06 00 00 00 00 0a 0
01/22/09 15:41:19: EVT#12860-01/22/09 15:41:19: 113=Unexpected sense: PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 15 11 00 00 10 00, Sense: 70 00 05 00 00 00 00 0a 00 00 00 00 2
01/22/09 15:41:19: EVT#12861-01/22/09 15:41:19: 113=Unexpected sense: PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 15 11 00 00 10 00, Sense: 70 00 05 00 00 00 00 0a 00 00 00 00 2
01/22/09 15:41:19: EVT#12862-01/22/09 15:41:19: 113=Unexpected sense: PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 1a 08 19 00 ff 00, Sense: 70 00 05 00 00 00 00 0a 00 00 00 00 2
---/errors---

In addition to the above it would eventually throw messages telling me the array was DEAD and unrepairable when connected over the expander, but replugging the drives directly to the controller and all was good (bar a rebuild that is - which I overrode via the cli).

I have seen reports like this before where SAS controllers would not play nice with SATA drives, but considering it's working great connected directly to the card...

Do let me know what you find out though (if the errors are similar to those above).
post #1718 of 7891
Hmm, that's not very promising, regarding the errors connected to the chenbro SAS expander. How long are the cable lengths? Same multilane breakout cables vs direct connect? Did you try the external port instead, if you have the relevant cables?

I'll put in my own testing with my Adaptec cards and the 28 port chenbro once all my parts arrive, probably this weekend. Then we'll know whether we have a viable solution for drive expansion, or a $200-300 brick...
post #1719 of 7891
Quote:
Originally Posted by kapone View Post

huh? It supports up to full e-atx motherboards.

I may have spoken too soon...Since it uses a standard ATX PSU (on it's side), the layout may not support ATX or higher motherboards. Hmm...That case does look nice. I may order one just to play with.

Norco page:
http://www.norcotek.com/item_detail....delno=rpc-3116

They have a sample pic there, with a server-sized motherboard installed.
It looks like the PSU just "floats" above the extra motherboard area beyond the first 4 slots. So you should still be able to use normal ATX or eATX motherboards, but the last 3 slots will be blocked by the PSU.
post #1720 of 7891
Quote:
Originally Posted by emigrating View Post

Is that in relation to the PD xx errors you posted earlier?

I've been getting all sorts of errors (see below) on some Seagate 1.5TB drives connected through the Chenbro CK13601 and eventually the controller would give up and die on me requiring a reboot to come back to life. Pulled the 1.5TB drives from the expander and connected them straight to the controller and all is good. I'm gonna try to reflash the RAID card now and see if that makes a difference as well as try some smaller 500GB drives.

---errors---
01/22/09 15:40:48: Task Mgmt Start Addr a14d3400 DevId[1d] Index 0 chip 0 type 3 numCmdIssued 9
01/22/09 15:40:48: EVT#12829-01/22/09 15:40:48: 267=Command timeout on PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 2a 00 06 51 41 80 00 00 80 00
01/22/09 15:40:48: EVT#12830-01/22/09 15:40:48: 267=Command timeout on PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 2a 00 06 51 41 00 00 00 80 00
01/22/09 15:40:48: EVT#12831-01/22/09 15:40:48: 267=Command timeout on PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 2a 00 06 51 42 80 00 00 80 00
01/22/09 15:40:48: EVT#12832-01/22/09 15:40:48: 267=Command timeout on PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 2a 00 06 51 43 00 00 00 80 00
01/22/09 15:40:48: EVT#12833-01/22/09 15:40:48: 267=Command timeout on PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 2a 00 06 51 40 00 00 00 80 00
01/22/09 15:40:48: EVT#12834-01/22/09 15:40:48: 267=Command timeout on PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 2a 00 06 51 42 00 00 00 80 00
01/22/09 15:40:48: EVT#12835-01/22/09 15:40:48: 267=Command timeout on PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 2a 00 06 51 40 80 00 00 80 00
01/22/09 15:40:48: EVT#12836-01/22/09 15:40:48: 267=Command timeout on PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 2a 00 06 51 43 80 00 00 80 00
01/22/09 15:40:48: EVT#12837-01/22/09 15:40:48: 268=PD 1d(e0xff/s29) Path 5001c4500000f50c reset (Type 03)
01/22/09 15:40:48: MPI_EVENT_SAS_DISCOVERY: PortBitmap ff - Discovery is in progress
01/22/09 15:41:18: Disc-prog= 0....resetProg=0 aenCount=0 transit=0
01/22/09 15:41:19: EVT#12859-01/22/09 15:41:19: 113=Unexpected sense: PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 2a 00 06 65 f2 80 00 00 80 00, Sense: 70 00 06 00 00 00 00 0a 0
01/22/09 15:41:19: EVT#12860-01/22/09 15:41:19: 113=Unexpected sense: PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 15 11 00 00 10 00, Sense: 70 00 05 00 00 00 00 0a 00 00 00 00 2
01/22/09 15:41:19: EVT#12861-01/22/09 15:41:19: 113=Unexpected sense: PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 15 11 00 00 10 00, Sense: 70 00 05 00 00 00 00 0a 00 00 00 00 2
01/22/09 15:41:19: EVT#12862-01/22/09 15:41:19: 113=Unexpected sense: PD 1d(e0xff/s29) Path 5001c4500000f50c , CDB: 1a 08 19 00 ff 00, Sense: 70 00 05 00 00 00 00 0a 00 00 00 00 2
---/errors---

In addition to the above it would eventually throw messages telling me the array was DEAD and unrepairable when connected over the expander, but replugging the drives directly to the controller and all was good (bar a rebuild that is - which I overrode via the cli).

I have seen reports like this before where SAS controllers would not play nice with SATA drives, but considering it's working great connected directly to the card...

Do let me know what you find out though (if the errors are similar to those above).

Yes, it is in relation to the errors I had previously mentioned... and your errors are similar to the errors I've been seeing (ie, Unexpected sense, Command timeout, PD reset -- I take it you aren't using the MagaRAID Storage Manager?).

Right now, I'm testing for a backplane for issues as I've pulled the drives out of the chassis (pain in the arse) and have directly connected them to the Expander. I think I'll directly connect to HBA like you are doing and see what that does.
post #1721 of 7891
I'm using the same cables, the only difference being they are plugged into the Expander instead of the HBA. The cable between the HBA and Expander is somewhat long at 60cm but shouldn't cause a problem. I did also stick some none RAIDed drives onto the expander earlier and they reported no sense errors so I believe it's something to do with the 1.5TB's rather than a SATA issue as such.

Anyhow, just reflashed the controller (I did find some CDB 15 and a1 references in the firmware readme which looks promising) and will give it another try after I've watched a couple eps of 24


plasticquart: Ok, I also had a bunch of Power ON, Reset, or Bus Device Reset occured errors.
Yes, I am using the storage manager, but it was easier getting at the logfile from MegaCLI

The firmware I grabbed (which mentioned some of the sense errors above) was just a day or so old, so could be worth you checking if you're up to date also. I'll do some more testing soon though.
post #1722 of 7891
Yeah so you have a wasted 3 expansion slots but it can still hold a full size motherbaord. Kind of sucks if you ask me. But I guess most wouldn't use all the expansion slots anyways? Depending.
post #1723 of 7891
Aren't those Seagate 1.5TB drives known for having issues?
post #1724 of 7891
Quote:
Originally Posted by kapone View Post

Did you make the tcp global changes on Server 2008 as well?

Yup. They are all set the same.
post #1725 of 7891
Quote:
Originally Posted by MiBz View Post

Mike, could you check on your 2008 server, in the physical nic settings if Large Send Offload v2 (IPv4) is enabled.

If you have Hyper-V installed, also check the virtual switch nic settings if
TCP Large Send Offload (IPv4) is enabled.

Yes on both counts. It's an Intel Gigabit ethernet controller (PCI-E) which in my experience has been very fast and reliable.
post #1726 of 7891
Quote:
Originally Posted by MikeSM View Post

Yes on both counts. It's an Intel Gigabit ethernet controller (PCI-E) which in my experience has been very fast and reliable.

Edit: sorry should have been more clear ...doing too many things at once.

In the Virtual adapter properties choose configure,

then set these to disabled

IP Checksum Offload (IPv4)
TCP Checksum Offload (IPv4)
TCP Large Send Offload (IPv4)

then try testing file transfer speed again
post #1727 of 7891
Quote:
Originally Posted by Dougie085 View Post

Aren't those Seagate 1.5TB drives known for having issues?

Yes, they are. Supposedly they released a fixed firmware just today, but it only goes up to the 1 TB drives. I guess they are still looking at the 1.5 TB fix? You can read about the saga here.

If you already have those drives it might be worth going to Seagate and downloading the new FW. They normally make you call and beg for FW updates, but there are so many people out there with problem drives you can just download it from their website. I think it is called SD1A-2B. At least that is what the file is called. It is not clear if this is the same 'SD1A' FW they have been providing for awhile now.
post #1728 of 7891
Quote:
Originally Posted by Dougie085 View Post

Aren't those Seagate 1.5TB drives known for having issues?

Yes and no.

Mine are the "fixed" firmware and have been performing beautifully running directly from the HBA. Since other people are seeing similar errors with "good" drives I don't think it's a drive issue as such and more likely a HBA firmware issue or general SATA incompability. I'm about to do some more testing now.


Quote:
Originally Posted by lifespeed View Post

Yes, they are. Supposedly they released a fixed firmware just today, but it only goes up to the 1 TB drives. I guess they are still looking at the 1.5 TB fix? You can read about the saga here.

The 1.5TB drives has had "fixed" (I say "fixed" because we still don't know if it's prone to suddenly disappear from BIOS, but that's not the isse I've seen today nor is it what the 1500's were well known for initially) firmware available for quite some time already. Both SD1 and CC1H "models" are supposedly fine, at least mine have been performing great in and out of RAID (apart from this issue today which I think was caused by the Expander).


---edit---
Ok, so the HBA firmware update don't seem to have helped matters much. I haven't seen any of the [previous] Unexpected sense or Command timeout errors this time around but it's now giving me Information unit CRC error detected on two of the drives instead. This is still on the 1.5TB drives. The single 500GB drive is performing fine, but I'm about to create a R6 of 500GB drives to see if the RAID level has anything to do with it.
post #1729 of 7891
Supposedly fixed I would say Who knows really. Could still be a drive issue.
post #1730 of 7891
Quote:
Originally Posted by emigrating View Post

Yes and no.

Mine are the "fixed" firmware and have been performing beautifully running directly from the HBA. Since other people are seeing similar errors with "good" drives I don't think it's a drive issue as such and more likely a HBA firmware issue or general SATA incompability. I'm about to do some more testing now.



The 1.5TB drives has had "fixed" (I say "fixed" because we still don't know if it's prone to suddenly disappear from BIOS, but that's not the isse I've seen today nor is it what the 1500's were well known for initially) firmware available for quite some time already. Both SD1 and CC1H "models" are supposedly fine, at least mine have been performing great in and out of RAID (apart from this issue today which I think was caused by the Expander).


---edit---
Ok, so the HBA firmware update don't seem to have helped matters much. I haven't seen any of the [previous] Unexpected sense or Command timeout errors this time around but it's now giving me Information unit CRC error detected on two of the drives instead. This is still on the 1.5TB drives. The single 500GB drive is performing fine, but I'm about to create a R6 of 500GB drives to see if the RAID level has anything to do with it.

Hmm, that kind of CRC error is usually associated with a cable problem isn't it?
post #1731 of 7891
Quote:
Originally Posted by MikeSM View Post

Hmm, that kind of CRC error is usually associated with a cable problem isn't it?

Yeah, both the CRC failure and the Bus Reset errors would indicate bad cables and/or backplanes - however the only difference between my setup with and without the Expander is ... the expander itself.

I'm using the exact same fanout cables to the drives, I've tried both input ports on the expander and all the different output ports.

I even opened another set of cables I had laying about so I've doublechecked all the fanout cables as well as the HBA -> Expander cable. Think that about covers all my bases apart from the Expander itself being bad which I somehow doubt as at least another member here is seeing similar errormessages.

I've emailed LSI, Supermicro and Chenbro now - hopefully one of them will get back to me with some info. Guess I should include Seagate on that list also, but I'm gonna try some other drives first, just have to finish shuffling the data off of them.

---edit---
So, Supermicro got back to me via email already, saying;
Quote:


I was seeing the specifications of the Chenbro Expander board and I did
not see any details on support for sata hdds. I am also saying this
based on our experience with expanders that the expander chip does not
go well with sata hdds. Can you confirm that your config of using sata
drives would work with the chenbro board by contacting them. Please let
me know your feedback on this.

This does fit in with what I've come across earlier, however I have another very similar setup that's working correctly. I've emailed Chenbro asking them to verify SATA compability though (even though they are using a LSI Expander IC which claims - afaik - SATA support and the Chenbro manual does state SAS/SATA compability).

Chenbro also got back to me, but so far they were rather useless as they were just pointing out;
Quote:


Dear Sir,
For a best performance, below are recommanded RAID card model for very kindly reference.
1. Intel: SRCSASRB
2. LSI: MegaRAID SAS 8344ELP/8308ELP/84016E/8708ELP/8888ELP

So they obviously didn't read my entire support email. Obviously I emailed them back immediately asking for clarification as I am indeed using one of the "supported" LSI controllers (well sort of, the LSI IC on my board is the same as one of the supported ones - of course I didn't tell them I was using a OEM product ).

An interesting find though, I moved the firmware/bios jumper (it's now on the "old bios" setting) on the Expander and that seems to have helped somewhat. I'm still getting a bunch of errors, but read/write speed is tenfold what it was previous. It's still not fast by any means, with reads around 300MB/s and writes around 100MB/s (compared to, I believe, roughly 700MB/s RW when connected directly to the HBA) but it's usable - although the CRC errors are bothering me as I have a feeling it may cause issues should one of the drives actually go bad.

---edit 2---
Chenbro got back to me again and after a lot of mailing back and forth this is the reply I got;
Quote:


Thanks for your reply and information , after we have duplicated this issue again by testing in house , we got below result for your reference :

1. testing configuration :

- RAID card : LSI MegaRAID SAS 84016E w/ latest firmware version : 1.12.220-0560
- Chenbro SAS expander card : CK13601 rev A1
- SATA II HDD : we tested two models from Seagate
a. Barracuda 7200.10 (ST3750640AS) , SATAII 750GB
b. Barracuda 7200.11 (ST31000340NS) , SATAII 1TB
- Chenbro SAS/SATA II backplane ( 4 ports , for Chenbro rackmount chassis )

2. testing software : I/O mega

3. testing result :
for four ports , total performance can be up to rough 240 ~ 280 MB/s for each HDD model , thus , each HDD performance can be up to 60 ~ 70 MB/s (normal is around 50 MB/s) , though the whole system is still working fine without performance downgrade , we still get the similar WARNING MESSAGE from RAID card log file as well , but we think it is just warning instead of Error message to force HDD off line, we think it is kind of communication issue between expander card and RAID card , supposedly, it is not big issue for the system


4. check point for you :
- make sure the firmware setting jumper shall be on pin 1 and 2
- check LSI Mega RAID firmware on hand , and update if necessary
- check CK13601 w/Revision A1

So at least they are acknowledging the problem. Now to see if we can come up with an actual fix I did some more playing around with settings yesterday and I'm now getting reads of around 700MB/s and instead of the flood of errors in the log I'm getting one or two every few minutes. Write speeds are still somewhat slow at 300-400MB/s (but it could be when I said RW 700MB/s that was the numbers for my 3Ware controller - can't be bothered to reconnect the drives directly to the HBA to do another test just now).
post #1732 of 7891
Anyone has tried iSCSI Subsystem from Addonics?
post #1733 of 7891
Quote:
Originally Posted by MikeSM View Post

BTW, the transfers seem capped to 35 MBytes/sec in both directions.

I have an Atheros integrated INC in one of my client machines. I don't care enough to troubleshoot it, but I used to get 60 MB/sec with an older set of drivers. I got newer drivers via Windows Update and now get no more than 35 MB/sec.
post #1734 of 7891
New Firmware SD1B for 1.5TB seagates:
http://seagate.custkb.com/seagate/cr...p?DocId=207957

I'll try them out when I get home...
post #1735 of 7891
Quote:
Originally Posted by dj9 View Post

I have an Atheros integrated INC in one of my client machines. I don't care enough to troubleshoot it, but I used to get 60 MB/sec with an older set of drivers. I got newer drivers via Windows Update and now get no more than 35 MB/sec.

Interesting. Which drivers worked fine. I have a hard time believing it's drivers since netperf can saturate the link just fine. But I'm happy to try new drivers if I can figure out which ones to test.

Edit: Updating to the latest drivers (12/16/08) from the atheros website doesn't change anything.


thx
mike
post #1736 of 7891
Quote:
Originally Posted by MiBz View Post

Edit: sorry should have been more clear ...doing too many things at once.

In the Virtual adapter properties choose configure,

then set these to disabled

IP Checksum Offload (IPv4)
TCP Checksum Offload (IPv4)
TCP Large Send Offload (IPv4)

then try testing file transfer speed again

Just tried this again. No difference.
post #1737 of 7891
Quick update the my previous posts...

Recap: Setup is an LSI MegaRAID 84106E connected to a 36-port SAS Expander (using mini-SAS to mini-SAS cable) which is connected to my 4020's backplane (using 4xSATA to mini-SAS (SFF8087) fan out cables) and the SATA drives are connected to the backplane.

Initial testing on a 4x1TB RAID-5 array was throwing errors in the MegaRAID Storage software. Suggestions from LSI was to simply the build and run tests to ferret out the issue... so I started by removing the Expander from the config (84016E connected to the backplane connected to the drives) tests ran still produced errors.

So I then went to the most basic config by removing the drives from the 4020 and directly connecting them to the MegaRAID 84106E card (removing the Expander and the backplane from the configuration). Testing showed the array to be solid with no errors being reported and great read/write speeds. Nice. Now I've reconnected the drives to the Expander (backplane is still out of the config) and have just started testing and so far no errors are being reported, but I just started this test. We'll see how it finishes.

Obviously it is looking like the backplane might be bad. I think I remember reading in here that someone else had a bad backplane in their 4020 case.

I'll try to get this testing finished up this weekend, but I don't think I'll have all the time I need. Either way, I'll followup on this once I work my way through it.
post #1738 of 7891
Quote:
Originally Posted by plasticquart View Post

Quick update the my previous posts...

Recap: Setup is an LSI MegaRAID 84106E connected to a 36-port SAS Expander (using mini-SAS to mini-SAS cable) which is connected to my 4020's backplane (using 4xSATA to mini-SAS (SFF8087) fan out cables) and the SATA drives are connected to the backplane.

Initial testing on a 4x1TB RAID-5 array was throwing errors in the MegaRAID Storage software. Suggestions from LSI was to simply the build and run tests to ferret out the issue... so I started by removing the Expander from the config (84016E connected to the backplane connected to the drives) tests ran still produced errors.

So I then went to the most basic config by removing the drives from the 4020 and directly connecting them to the MegaRAID 84106E card (removing the Expander and the backplane from the configuration). Testing showed the array to be solid with no errors being reported and great read/write speeds. Nice. Now I've reconnected the drives to the Expander (backplane is still out of the config) and have just started testing and so far no errors are being reported, but I just started this test. We'll see how it finishes.

Obviously it is looking like the backplane might be bad. I think I remember reading in here that someone else had a bad backplane in their 4020 case.

I'll try to get this testing finished up this weekend, but I don't think I'll have all the time I need. Either way, I'll followup on this once I work my way through it.

I was the one with the bad backplane. Is yours a V3.0 backplane? (the one with SAS and SATA connectors)? The norco guys were great about sending me a replacement. That new backplane is showing no CRC errors on it.

Thx
mike
post #1739 of 7891
Quote:
Originally Posted by MikeSM View Post

I was the one with the bad backplane. Is yours a V3.0 backplane? (the one with SAS and SATA connectors)? The norco guys were great about sending me a replacement. That new backplane is showing no CRC errors on it.

Thx
mike

No, mine only has SATA connectors on it. Did you get your replacement directly from Norco?

And looking at my backplane, it looks like it is composed of several independent PCB "strips" (if that makes sense) and I think there is slight physical damage to 3 of them.
post #1740 of 7891
Quote:
Originally Posted by plasticquart View Post

No, mine only has SATA connectors on it. Did you get your replacement directly from Norco?

And looking at my backplane, it looks like it is composed of several independent PCB "strips" (if that makes sense) and I think there is slight physical damage to 3 of them.

Yes. There was no obvious damage to any of the boards, but something happened along the way electrically I think.

Be careful when reassembling them to put back all the screws so the drives seat properly.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Home Theater Computers
AVS › AVS Forum › Video Components › Home Theater Computers › Guide To Building A Media Storage Server