| 1 |
On Thursday 30 November 2006 09:02, Etaoin Shrdlu wrote: |
| 2 |
> On Thursday 30 November 2006 07:56, Vladimir G. Ivanovic wrote: |
| 3 |
> > Let's take a poll. |
| 4 |
> > |
| 5 |
> > 1. Have you seen this error message in an emerge? |
| 6 |
> |
| 7 |
> Yes, several times. |
| 8 |
|
| 9 |
Ditto. |
| 10 |
|
| 11 |
> > 2. Have you subsequently identified a hardware problem, fixed the |
| 12 |
> > hardware problem, and have not seen the message since? |
| 13 |
> |
| 14 |
> Yes. 99% of the times it was bad RAM (verified with memtest86). |
| 15 |
> Of course, for trivial emerges a subsequent emerge completed fine, but |
| 16 |
> the first failure put me on the alert. |
| 17 |
|
| 18 |
Yes, it was a bad/incompatible RAM module for the particular memory |
| 19 |
controller. memetest86 did not identify it, I found out by trial and error! |
| 20 |
The symptom was random crashes (mostly) during emerge which would complete |
| 21 |
fine after a hard reboot. The crashes would invariably happen either when |
| 22 |
the memory controller was switching onto the next memory module, or when both |
| 23 |
modules were used up and it started using swap. |
| 24 |
|
| 25 |
> > 3. Have you re-run the emerge and not seen the message in a while |
| 26 |
> > (please indicate how long "a while" is.) |
| 27 |
> |
| 28 |
> For me, "a while" is "since fixing the hardware problem". |
| 29 |
|
| 30 |
Ditto, i.e. about 18 months so far. |
| 31 |
|
| 32 |
HTH. |
| 33 |
-- |
| 34 |
Regards, |
| 35 |
Mick |
Archives