I have experienced a very strange problem last night that ultimately resulted in my SCIS Zip drive going away. This problem happened actually while I was working with Mac OS X Server, attempting to help Carl with his OS X <-> Linux mount problems. Here's what happened: I have a Mac 8600 which dual boots Mac OS 8.5 and Mac OS X Server. This is connected to my SuSE box running 7.2 / 2.4.4. The SuSE box acts as router / firewall to my DSL / The World. This all just works under OS 8.5 running Netatalk; I had to re-configure TCP/IP on the OS X side because I haven't been using it since I relocated -- it still had my old net setup with references to my former ISP. This got cleaned up in short order, and SuSE merrily forwarded traffic from OS X. The problem came when I fired up the OmniWeb web browser on OS X. Initially, it was trying to proxy through some IP I didn't immediately recognize. I disable this in OmniWeb, and was able to pull up sites. But I wanted to track down that proxy IP, so I tried to pull up my old ISP's website. The instant I fired off OmniWeb to get the ISP's page, SuSE went down Real Hard. Froze completely, couldn't get to other consoles on SuSE, couldn't shut down Xfree, couldn't telnet in. When I reset / rebooted the SuSE box, I noticed some odd SCSI parity errors in the boot messages, and I noticed that my SCSI Zip Drive (which hangs off of an Adaptec 29160N along with a Seagate 9gig harddrive) was gone. Then when I rebooted again, I noticed that the 29160 wasn't finding the Zip Drive -- Its Gone! I've tried to power down both the CPU and the ZIP, but the 29160 still doesn't see it. It sees the Seagate just fine, but not the ZIP. Any Clues as to where my ZIP drive is? Incidently, hitting my old ISP's site from Mac OS X will consistantly bring the SuSE box down. Every time. I can hit other sites just fine, but the ISP's site apparently is doom. Further, I can hit the ISP from my SuSE box without any ill effect, it only happens when I hit it from Mac OS X. (Not my month for SCSI...) -- --Gregory
The 02.10.18 at 13:04, Gregory Sawyer wrote: [...]
Incidently, hitting my old ISP's site from Mac OS X will consistantly bring the SuSE box down. Every time. I can hit other sites just fine, but the ISP's site apparently is doom. Further, I can hit the ISP from my SuSE box without any ill effect, it only happens when I hit it from Mac OS X.
Try to do it while the suse is displaying console-10, you know, the one with error logs. It might be able to report something there. Also, have a look at /var/log/firewall - it is the file kernel messages go to, but it could show nothing, depending on how bad is the crash. As to the dissapearing disk, I don't know :-? -- Cheers, Carlos Robinson
On Friday 18 October 2002 03:11 pm, Carlos Robinson wrote:
The 02.10.18 at 13:04, Gregory Sawyer wrote:
[...]
Incidently, hitting my old ISP's site from Mac OS X will consistantly bring the SuSE box down. Every time. I can hit other sites just fine, but the ISP's site apparently is doom. Further, I can hit the ISP from my SuSE box without any ill effect, it only happens when I hit it from Mac OS X.
Try to do it while the suse is displaying console-10, you know, the one with error logs. It might be able to report something there. Also, have a look at /var/log/firewall - it is the file kernel messages go to, but it could show nothing, depending on how bad is the crash.
As to the dissapearing disk, I don't know :-?
Thanks Carlos, for the suggestion - I'll try the console-10 lurk later. I did peek at the logs last night but couldn't find anything -- apparently the loggers weren't able to flush to disk. But I did just grep through /var/log/firewall, and I discovered that the ZIP drive was throwing nasty parity errors every morning at around 00:15, which is coincident with a cron script that I know runs at that time. That would be also about the same time this morning that my ex-ISP's site first crashed the machine. These started on the morning of the 16th -- on the 15th I compiled and installed the latest release of CUPS along with the CUPS folk's version of GhostScript -- and CUPS has taken the system down a couple of times as well in a similar manner. Just one more thing to convolute this insanity. -- --Gregory
The 02.10.18 at 17:44, Gregory Sawyer wrote:
Incidently, hitting my old ISP's site from Mac OS X will consistantly bring the SuSE box down. Every time. I can hit other sites just fine, but the ISP's site apparently is doom. Further, I can hit the ISP from my SuSE box without any ill effect, it only happens when I hit it from Mac OS X.
Try to do it while the suse is displaying console-10, you know, the one with error logs. It might be able to report something there. Also, have a look at /var/log/firewall - it is the file kernel messages go to, but it could show nothing, depending on how bad is the crash.
As to the dissapearing disk, I don't know :-?
Thanks Carlos, for the suggestion - I'll try the console-10 lurk later. I did peek at the logs last night but couldn't find anything -- apparently the loggers weren't able to flush to disk.
But I did just grep through /var/log/firewall, and I discovered that the ZIP drive was throwing nasty parity errors every morning at around 00:15, which is coincident with a cron script that I know runs at that time. That would be also about the same time this morning that my ex-ISP's site first crashed the machine.
Too many write errors could bring the system down, it happened to me. It tries to write something, but it can't; if the problem is the root (/) it fails again when trying to log that... and anyways, the system is too busy trying to cope: it seems to be unable to do anything else.
These started on the morning of the 16th -- on the 15th I compiled and installed the latest release of CUPS along with the CUPS folk's version of GhostScript -- and CUPS has taken the system down a couple of times as well in a similar manner. Just one more thing to convolute this insanity.
:-( -- Cheers, Carlos Robinson
participants (2)
-
Carlos E. R.
-
Gregory Sawyer