1 |
Hi Harry, |
2 |
|
3 |
Harry Putnam wrote: |
4 |
|
5 |
> Alan McKinnon <alan.mckinnon@×××××.com> writes: |
6 |
> |
7 |
>> On Wednesday 17 February 2010 00:36:42 Harry Putnam wrote: |
8 |
>>> Hundreds, maybe thousands of lines like this (wrapped for mail): |
9 |
>>> |
10 |
>>> Feb 16 09:38:47 reader kernel: [162289.090685] usb 4-2.1:1.1: uevent |
11 |
>>> |
12 |
>>> Feb 16 09:38:48 reader kernel: [162289.467065] hdc: status error: |
13 |
>>> status=0x00 { } |
14 |
>>> |
15 |
>>> Feb 16 09:38:48 reader kernel: [162289.467071] hdc: possibly failed |
16 |
>>> opcode: 0xa0 |
17 |
>>> |
18 |
>>> Feb 16 09:38:48 reader kernel: [162289.467079] ide-atapi: hdc: |
19 |
>>> Strange, packet command initiated yet DRQ isn't asserted |
20 |
>>> |
21 |
>>> When I noticed this output involving the cdrom I wondered if I might |
22 |
>>> have left something in it but that was not the case. |
23 |
>> |
24 |
>> Do you have hal configured to poll your cdrom drive every two seconds, to |
25 |
>> see if a disk is inserted? And if so, is the verbosity logging cranked up |
26 |
>> way higher than it should be? |
27 |
>> |
28 |
>> I haven't personally had to fix this myself (so can't give pointers on |
29 |
>> where to fix it), but it seems to be a common occurrence judging from |
30 |
>> posts I see here and at other forums. |
31 |
> |
32 |
> I do have hald running, but made no special config regarding cdrom |
33 |
> polling. At least not on purpose. |
34 |
> |
35 |
> The messages do appear to be continuous. I will execute a reboot soon |
36 |
> but don't want to right now. |
37 |
> |
38 |
> Why I'm pondering and following this up, is that I experience a |
39 |
> serious freeze after some unspecified amount of uptime. Mouse and |
40 |
> keyboard become unresponsive... and eventually the OS cannot be |
41 |
> accessed at all. |
42 |
> |
43 |
> SSH appears to stop and cannot contact remotely either. |
44 |
> |
45 |
> This began happening quite some time ago... on a different earlier |
46 |
> install. I never could see anything in the logs that gave a clue to |
47 |
> why. |
48 |
> |
49 |
> I created a script that ran from cron. It pinged a remote host, and |
50 |
> logged a unique easily findable string into the log using `logger', |
51 |
> every 5 minutes. With that I was able to narrow down the time frame |
52 |
> of freeze to within the last 5 minutes (of log lines). |
53 |
> |
54 |
> Even then, there was nothing to indicate a problem. This was an OS |
55 |
> that had been running a very long time with upgrade after upgrade. |
56 |
> |
57 |
> Though I hated having to rebuild all the customizations etc, I finally |
58 |
> completely reinstalled from scratch hoping to catch the problem with |
59 |
> the shotgun approach. |
60 |
> |
61 |
> In that earlier OS there were no log messages regarding hdc being |
62 |
> generated (by the way). |
63 |
> |
64 |
> Shortly after completing the new install and a couple of weeks of |
65 |
> getting setup the way I wanted, I began to experience the freezes |
66 |
> again. |
67 |
> |
68 |
> I have caught the freeze in the early stages before completely losing |
69 |
> the network when just mouse and keyboard became unresponsive, was able |
70 |
> to ssh in and noticed that restarting hald held off the freeze for |
71 |
> some (again unspecified) amount of time. |
72 |
> |
73 |
> So cutting the lengthy narrative down a bit, and briefly put, I'm |
74 |
> looking for anything unusual that is causing this. The hdc messages |
75 |
> is the only odd thing I'm seeing. |
76 |
> |
77 |
> Something appears to be jamming up the hal layer somehow, but not |
78 |
> leaving findable tracks. At least not findable by an someone with |
79 |
> many yrs experience with linux but not much real debugging of |
80 |
> complicated problems under his belt. |
81 |
|
82 |
I had once similar freezes and effects until I recognized that our rabbit |
83 |
had bitten into an USB cable and the computer got undefined signals on the |
84 |
USB due to contacts of the blank cable lines. |
85 |
|
86 |
Try to disconnect any external USB device first and check if the problem |
87 |
persists. |
88 |
|
89 |
- Jörg |