Fixy computer 1 dorra

[suityou01]

I spent all day yesterday at one of my clients. 'Twas a simple job. Pop some more ram in his server. 20 mins. I even wasn't sure what ram to get, so posted a query in technical about it. Ensuring I was armed with the correct information I orderd the ram, it turned up in the week, and I fitted it yesterday.

Fortunately clientco is a mate of mine, and soon to become my accountant. I fitted the ram, and it didn't firkin work. Beep beep beep, incorrect memory config please swap the ram from DIMM_1_A and DIMM_2_A to DIMM_2_A and DIMM_2_B.

Duly swapped.

Then, beep beep beep beep beep beep.......

Black screen, no post test.

Put the ram back to the original config, and take mine out.

Server does a post test. Then gives a warning about data loss and fails to boot.

I quickly figured out that the RAID controller wasn't starting (by the absence of the BIOS screen)

Then an overwhelming sense of panic set in.

Matey went to get some bacon rolls, while I nipped out for a crafty fag and gave this some thought.

What changed? I poked around inside the machine, then the raid controller stopped working. Why?

I calmly decided to reseat the raid controller, even though it is a daughter board which could not have become dislodged by gently pressing in some ram chips.

Raid controller reseated I pressed the on button with my heart in my mouth.

The frakkin thing booted first time, just as the bacon rolls arrived.

My mate knew I was worried, and I lost some face.

Moral of the story, don't buy dell servers, and if you do, don't let me anywhere near it.

HTH

[Durbs]

posted a query in technical about it.

There's the root cause, right there.

[stek]

Oddly enough, I'm upgrading RAM on a Sparc box in Watford. I'm in Leeds and it did this...

0:0:0>MSG = CE DRAM error,
MCU 0
CE nib:bits 141
DRAM AFAR 00000000.00000000
L2 AFAR 00000008.001b0820
AFSR 60000000.00007106
0:0:0>END_WARNING

0:0:0>Decode of Dram Error Log Reg Channel 0 bits 60000000.00007106
0:0:0> 1 MEC 62 R/W1C Multiple corrected errors, one or more CE not logged
0:0:0> 1 DAC 61 R/W1C Set to 1 if the error was a DRAM access CE
0:0:0> 7106 SYND 15:0 RW ECC syndrome.
0:0:0> Dram Error AFAR channel 0 = 00000000.00000000
0:0:0> L2 AFAR channel 0 = 00000008.001b0820
0:0:0> DRAM Addr for ROW = 0x00000001 COL = 0x00000586 BANK = 0x00000006
0:0:0>
0:0:0>ERROR: Correctable Dram error threshold exceeded.
0:0:0> 1 CEs detected at nibble:mask 00000000.00000141 channel 0.
0:0:0> First failing address = 00000008.001b0820.
0:0:0> Disabling L2 CEEN traps
0:0:0>
0:0:0>ERROR: TEST = Data Bitwalk
0:0:0>H/W under test = MB/CMP0/CH0/R1/D0/S0 (J0801)
0:0:0>Repair Instructions: Replace items in order listed by 'H/W under test' above.
0:0:0>MSG = Pin 81 failed on MB/CMP0/CH0/R1/D0/S0 (J0801)
0:0:0>END_ERROR

Got the local guy to reseat it and looks ok...

Phew, double time as well!

[suityou01]

Oddly enough, I'm upgrading RAM on a Sparc box in Watford. I'm in Leeds and it did this...

0:0:0>MSG = CE DRAM error,
MCU 0
CE nib:bits 141
DRAM AFAR 00000000.00000000
L2 AFAR 00000008.001b0820
AFSR 60000000.00007106
0:0:0>END_WARNING

0:0:0>Decode of Dram Error Log Reg Channel 0 bits 60000000.00007106
0:0:0> 1 MEC 62 R/W1C Multiple corrected errors, one or more CE not logged
0:0:0> 1 DAC 61 R/W1C Set to 1 if the error was a DRAM access CE
0:0:0> 7106 SYND 15:0 RW ECC syndrome.
0:0:0> Dram Error AFAR channel 0 = 00000000.00000000
0:0:0> L2 AFAR channel 0 = 00000008.001b0820
0:0:0> DRAM Addr for ROW = 0x00000001 COL = 0x00000586 BANK = 0x00000006
0:0:0>
0:0:0>ERROR: Correctable Dram error threshold exceeded.
0:0:0> 1 CEs detected at nibble:mask 00000000.00000141 channel 0.
0:0:0> First failing address = 00000008.001b0820.
0:0:0> Disabling L2 CEEN traps
0:0:0>
0:0:0>ERROR: TEST = Data Bitwalk
0:0:0>H/W under test = MB/CMP0/CH0/R1/D0/S0 (J0801)
0:0:0>Repair Instructions: Replace items in order listed by 'H/W under test' above.
0:0:0>MSG = Pin 81 failed on MB/CMP0/CH0/R1/D0/S0 (J0801)
0:0:0>END_ERROR

Got the local guy to reseat it and looks ok...

Phew, double time as well!

So all servers are bastids then.

I suspect that I bought non ECC ram, but that said, just putting the new ram in on it's own still didn't work. Right number of pins, right clock speed, numbers match perfectly

[DimPrawn]

I hope you've got REALLY good liability insurance.

This is why £15 fix businesses don't last.

One day you'll feck up someone's server and find a letter from their lawyers taking you to court for £5M worth of damages.

Enjoy!

[d000hg]

Moral of the story, don't let me anywhere near it, use a professional rather than a programmer/BA who knows how to take the case off.

FTFY. (only kidding)

[suityou01]

FTFY. (only kidding)

Doogie in the previous 15 years as a techy I forgot more than you're likely to learn, so go and get your "blanky" and have your mid morning nap ya young scamp.

[suityou01]

I hope you've got REALLY good liability insurance.

This is why £15 fix businesses don't last.

One day you'll feck up someone's server and find a letter from their lawyers taking you to court for £5M worth of damages.

Enjoy!

I really thought you were one of the brighter posters, comparatively speaking. The £15 fix is for saturday mornings only and only applys to domestic clients. Businesses pay full fare, although obviously yesterday I cannot charge for and I have to chalk up to experience.

Really hope this sinks in, if not PM me for scanned crayon drawings that break this down into even more simple concepts.

[stek]

I hope you've got REALLY good liability insurance.

This is why £15 fix businesses don't last.

One day you'll feck up someone's server and find a letter from their lawyers taking you to court for £5M worth of damages.

Enjoy!

I'm permie - I can break anything I want!

But let's say I'm not - and I bugger summer up, and I've no insurance, where's the £5m gonna come from? Personally I'm worth nowt, suppose I could pay it back over the rest of my life at £100,000 a month but that would mean taking a second job.....

[d000hg]

Doogie in the previous 15 years as a techy I forgot more than you're likely to learn, so go and get your "blanky" and have your mid morning nap ya young scamp.

I bow to your superior knowledge of obsolete hardware in that case. When my 386 running Dos 3 breaks, I'll give you a call


I thought most of us techies spent their time trying to stop friends and family assuming that because we "work in computers" we can do tech-support. Sssh!