On 23 Feb 2003 at 12:19, Michael Schwendt wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > On Sun, 23 Feb 2003 11:45:30 +0100, Stefan Neufeind wrote: > > > Since it's a server I don't run or even have a graphical surface > > installed. Hang means: Console (Textmode) doesn't respond anymore, > > can ping to machine okay, can connect to services but don't get a > > respond. Some people said it might be a problem with the harrdisk- > > layer. Assuming this would for sure make clear why the running > > kernel can work properly but neither can a service spawn new > > children (ssh, ftp, ...) nor could syslog write to disk (and maybe > > cause everything to stop) nor could the swap-space be accessed. > > You could configure sys(k)logd to write *all* kernel log messages to > console. Maybe that will turn something up. > > > Does anybody know about problems with harddisks? > > U-DMA related crashes as well as certain hard disks failing to > return from sleep mode, which effectively blocks the kernel and new > processes which want to access a disk. I have entered that into > bugzilla a long time ago, and it's still there, most likely because it > can also be due to a firmware bug that will never be fixed. IIRC, my > work-around is either to avoid sending the disk into sleep mode or to > switch off DMA. Due to a proposal from somebody else I changed from udma5 to udma4 now and turned off 32-bit-transfers to the default 16-bit mode. Sleep mode is not active, as far as I know, since it's a server system. I don't send the drives to sleep myself or anything like that. How can I determine if auto-spindown/-sleep is active? So you say either use sleep or dma, right? If I disable sleep I should be able to run udma? Maybe the problem is with udma5 and now gone because of switching to udma4 (I hope so!). There are no messages in syslog reporting harddisk-problems. From other posts I learned that often hardisk-problems are combined with messages in syslog indicating harddisk-problems. But there aren't any. > I haven't heard about other problems with harddisks. Hmm ... maybe it's not the harddisk and we have just ran into some deadend. But this is the only thing common for my two servers which both show the same symptoms. -- Psyche-list mailing list Psyche-list@redhat.com https://listman.redhat.com/mailman/listinfo/psyche-list
