Well, don't do firmware updates while sleep deprived... (CacheCade now inoperative)

[anomaly]

Seems like common sense but after hours of frustration with a 9361-8i card (legitimate card, I got it from eBay but it came with a proper retail box), I got it to work with two of my SAS SSDs for Cache Cade.

The card has a 8GB CV module, and Cache Cade key. Following Murphy's Law, I decided to check for updates to the firmware and installed storlci and co. After that, I grabbed the latest ZIP archive and began flashing. I came across a file "70-26832-00_Ptolemy4GB_GG_VPD.rom". I decided to go ahead and flash it after checking the CV module had a less recent number, and per the title, a nasty mess ensued. The two SAS SSDs that were working before, now exhibit this behavior:

---------------------------------------------------------------------------------
EID:Slt DID State DG Size Intf Med SED PI SeSz Model Sp Type
---------------------------------------------------------------------------------
252:0 9 UGood - 5.457 TB SATA HDD N N 512B HGST HDN726060ALE614 U -
252:1 10 UGood - 5.457 TB SATA HDD N N 512B HGST HDN726060ALE614 D -
252:2 11 UGood - 5.457 TB SATA HDD N N 512B HGST HDN726060ALE614 D -
252:3 8 UGood - 5.457 TB SATA HDD N N 512B HGST HDN726060ALE614 D -
252:4 12 UBad - 0 KB SAS HDD Y N 512B HUSMR1625ASS201 U -
252:5 16 UBad - 0 KB SAS HDD Y N 512B HUSMR1625ASS201 U -
252:6 14 UGUnsp - 0 KB SAS HDD N N 0 KB HSSC0480S5xnNMRI U -
252:7 15 UGUnsp - 0 KB SAS HDD N N 0 KB HUSMR1640ASS204 U -

I have tried everything from cold starts, to clearing all configuration settings, to switching to JBOD mode and back, and nothing brings back the 1625 HGST SAS SSDs. They remain UBad, and whenever I try anything they just forbid any chances to the state of the drives.

The worst thing is that they are reported as 0 size, but their sector information reports that properly (512, 520 for the 1640 and 0480S drives, they need reformatting).

I'm at a loss here, so I would appreciate any help. I already filed a ticket with LSI too.

For now the take away lesson is "don't mess with firmware while sleep deprived". But every now and then we screw up with things we usually never bother ourselves with.

 

[anomaly]

I'm still working on this with LSI now involved, they provided an image for the CV firmware, but after flashing that, now the option to use the Intel DC S3700 SSDs is also gone, almost like Cache Cade is not available. The issue with the drives is fixed: the controller had left them in an inconsistent state and I had to take them out, feed them into a SA120 and use sg3 utils to reformat them. The 520B sector drives are now 512B and working fine, or so it seems.

A sidenote: a word of caution on the Norco 4U 4408 cases: the PCB of the SAS 12Gbps backplane is garbage. I had to rework the entire thing in my electronics lab. The solder joints were cold, and they used the wrong kind of solder from the looks of it, relying on two tiny screws to hold the connector in place with pressure, but the solder joints were so brittle they eventually dislodged and it was causing intermittent issues with power to the upper row of drives. I will create another thread for this as a PSA of sorts... Trying to get Inter-Tech in Europe to do their job proper and send a replacement PCB but they want the original back (more downtime... you are selling IPC products and want a PCB worth 5 bucks in parts sent back across Europe before you send a replacement? what kind of nonsense is that when you have clear photos of what happened as part of faulty QA?).

Does anyone know why CacheCade could suddenly become unusable in the controller when it was working fine before the firmware that LSI provided me during support was flashed (they gave me the wrong 8GB CV module firmware it seems).

 

[i386]

Maybe you Had a Software Key for cachecade

 

[anomaly]

Nope, using hardware key. I managed to fix this, and will comment later in case anyone runs into the same hell.

I had to take the CV module out, do a factory defaults reset, boot, reflash firmware, shutdown, install CV module and then it worked.

 

[ecosse]

A sidenote: a word of caution on the Norco 4U 4408 cases: the PCB of the SAS 12Gbps backplane is garbage. I had to rework the entire thing in my electronics lab. The solder joints were cold, and they used the wrong kind of solder from the looks of it, relying on two tiny screws to hold the connector in place with pressure, but the solder joints were so brittle they eventually dislodged and it was causing intermittent issues with power to the upper row of drives. I will create another thread for this as a PSA of sorts... Trying to get Inter-Tech in Europe to do their job proper and send a replacement PCB but they want the original back (more downtime... you are selling IPC products and want a PCB worth 5 bucks in parts sent back across Europe before you send a replacement? what kind of nonsense is that when you have clear photos of what happened as part of faulty QA?).

Somewhat offtopic but I'd never buy a Norco produced enclosure again - well not if I want to keep my data intact anyway.

 

[anomaly]

Somewhat offtopic but I'd never buy a Norco produced enclosure again - well not if I want to keep my data intact anyway.

Yeah, the QA for the manufacturing process is a joke for the more complex components, like the backplanes. I had to rework the PCB for one of them because the molex 4-pin socket they use is poorly soldered and it dislodges eventually, causing erratic power connections between the PSU and the backplane. They used low quality sockets with surface 'spot weld' pins, and rely on two tiny screws on the plastic to hold it together. Needless to say, the solder they used did not flow proper because the metal alloy of the pins is crap, and they should not be using that connector for mechanically stressed connections at all. I ended up resoldering them with a Hakko station, applying urethane coating to the PCB lines, and used conformal coating underneath the socket for extra strength, then screwed it down onto the pins, and it works fine now.

They sell as Inter-Tech in Europe. I filed for a RMA for the PCB alone knowing the return they will send will likely have crappy connectors.... and guess what, they want the old, manually fixed one BACK. 6 USD worth of parts tops... shipped abroad. Yeah. Heh.