Drive Bay issues

[Msk]

Hi,

I have 2 HP G8 servers which each have 16 Drive Bays using LSI SAS 9211-8i controllers

Issue I am having is that on both servers freenas only detects 15 of the 16 drive bays

on both servers it doesnt detect drive bay 7.

Is this normal? I have replaced drive bays, cables LSi controllers and Harddrive and same thing happens everytime drive bay 7 is dead

 

[Ericloewe]

Are you using the correct firmware (P16 IT)?

 

[jgreco]

What does the BIOS see? Control-somethingorother will take you into that during POST, unless you've failed to flash the BIOS component. Some people advocate that, but in general I see it as foolish to omit unless you have a damn good reason for doing so, since it is useful for determining whether a problem is hardware or software in nature. Like this.

 

[mjws00]

My money is on a driver/kernal? glitch. The example that jumps to mind is @AltecBX 's nice E5 build. SM 836E26 chassis, MBD-X10SRH-CLN4F. FreeNAS would not enumerate lower left (first) drive bay. All other OS's tried worked correctly. All drives visible during post, FreeNAS misses that slot. Looks like the expander steals the first device somehow (ala dmesg).

We thought it might be the 3008 controller at the time, so it is interesting to see this pop up on a 9211-8 and HP gear. The error was discovered during burn-in around page 6. What was interesting also was @depasseg has the same mobo in the superserver chassis with the sas3 backplane without issue. Unfortunately that exact rig isn't terribly common for troubleshooting. But interesting to see a similar 16 bay chassis (same expander chip?) exhibiting a similar problem.

Good luck, AltecBX had the option of just skipping that bay and moved on. This is the next one we've seen.

 

[Ericloewe]

What does the BIOS see? Control-somethingorother will take you into that during POST, unless you've failed to flash the BIOS component. Some people advocate that, but in general I see it as foolish to omit unless you have a damn good reason for doing so, since it is useful for determining whether a problem is hardware or software in nature. Like this.

The P16 extension ROM apparently has trouble recognizing drives, which tends to defeat its whole purpose in IT mode.

 

[jgreco]

I've never seen that issue, and I suspect I have more LSI cards than most of the users here.

 

[Ericloewe]

I've never seen that issue, and I suspect I have more LSI cards than most of the users here.

You probably do, but it's in LSI's P17 changelog (or maybe the P16 known issues section of the readme). There's also the occasional thread asking if everything is alright despite the LSI extension ROM not showing the drives properly.

 

[jgreco]

I'd rather have the tool available, even if it is unreliable in some cases. I learn nothing if the BIOS is not there. With the BIOS, I stand the chance of learning that the BIOS correctly sees all the drives, which would clarify in this case if it is some sort of driver issue instead of a hardware problem like a bad connector.

It is not clear to me what value is gained by removing the BIOS. Well, I guess it makes sure that you can't be led astray by missing BIOS results ... but that's kind of weird logic.

 

[marbus90]

Since you have that bunch of card around, can you check if you can still enable staggered spinup? I know it might not work in most circumstances, but worth a shot.

Also, /which/ G8 server are you talking about. There are hundreds of G8 models around...

 

[jgreco]

Most consumer SATA drives do not support Staggered Spin-Up, and even if they do, they probably need to be on a backplane to take care of the signalling (pin 11). I don't have a backplane and sufficient SAS drives handy to test at the moment to figure out what the M1015 does. You can always bop into the BIOS and look if there's an option ;-)

 

[depasseg]

Regarding my and @AltecBX troubleshooting, my sense was it had something to do with the way the enclosure was being detected. In my case, the controller recognized the enclosure as an enclosure, but in his, it looked like the controller was interpreting the enclosure as one of the drives.