rbbb@RICE.ARPA (12/16/84)
It is still possible to get the tbuf errors with rev 7 hardware. We have always had rev 7 hardware, and we have never had the writeable control store. We once had lots of errors, and now we do not. How did we fix this, back in the days before microcode updates? I believe that we replaced the L0001 module with one that used a different brand of chips. This fix was discovered through a great deal of board-swapping; if anyone has a similar story, please let me know, because this all happened a looong time ago (but we still don't have the PCS, and we still run rev 7 hardware, and we don't get tbuf errors, and we swapped a pile of boards; I just can't remember exactly which one). And it is also true that Unix seems to get more of these than VMS (including the user-mode machine checks). Weird as shit. drc
sherouse@unc.UUCP (12/18/84)
In article <brl-tgr.6592> rbbb@RICE.ARPA writes: >It is still possible to get the tbuf errors with rev 7 hardware. We have >always had rev 7 hardware, and we have never had the writeable control >store. We once had lots of errors, and now we do not. How did we fix >this, back in the days before microcode updates? I believe that we >replaced the L0001 module with one that used a different brand of chips. >This fix was discovered through a great deal of board-swapping; if anyone >has a similar story... Similar is not the word. We had exactly the same experience. The original L0001 was *so* bad we couldn't keep the machine up long enough to read the distribution tape! We are now crashing about once every two weeks with a light load. Is it true that there is a *tested* and available version of the micropatch loader for 4.2bsd? And does it really help? George W. Sherouse Associate Physicist Division of Radiation Therapy North Carolina Memorial Hospital Chapel Hill, NC 27514 (919) 966-1101 <decvax!mcnc!unc!godot!sherouse>
sherouse@unc.UUCP (12/18/84)
I meant L0003, not L0001. rbbb@rice suckered me.