Re: [expert] K7S6A/SiS745+2.4.18+HIGHMEM freeze(news)
Same thing for me : I insert noapic in GRUB, but my gateway still hangs randomly. I'll try to recompile the 8.2 kernel, and/or to downgrade to the 8.1 kernel (hope that the 8.1 kernel will run fine with the 8.2 distribution). N. Le Vendredi 17 Mai 2002 04:14, vous avez écrit : hi: noapic didn't work for me. the server will hang under high loading after 4 or 5 hours. i use ctcs testing program to test. mdk 8.1 or red-hat 7.2 didn't have this problem. sorry i didn't have a chance to test it with red-hat 7.3, bc4 the server is one line with mdk 8.1 now... I should report that I got a message from Tom Wanka. The noapic boot option worked for him (he had a serverworks chipset), so this will probably also work for you. We do not use this chipset, so it is not clear if it will work in our case. There could also be a bug in the kernel. I found this interesting line in Alan Cox's description of changes in the 2.4.19 kernel: o Fix highmem truncation on DMA mapping bug (Dave Miller) I do not know if this bug was introduced after 2.4.18, or if it is in 2.4.18. Bjarne On Thu, 2002-05-16 at 08:48, [EMAIL PROTECTED] wrote: hi: i have same problem. 2.4.18 with serverworks and himem, the system hang. i report this several times. but no support or reply yet. Have you seen this! [RHBA-2002:085-11] Kernel panic on SMP systems with ext3 file systems is now fixed. I am not sure if this problem is solved in the Mandrake kernel. -- Bjarne On Wed, 2002-05-15 at 21:58, Bjarne Thomsen wrote: Actually it is on purpose that I did not mention VIA and the PCI latency problem in the subject line. I believe that we have two different problems (unfortunately). The K7S6A was a replacement for an Asus A7V266E with a VIA KT266A chipset. The reson was that we had similar lock-ups, often accompanied by massive disk corruptions which we have not seen with the K7S6A. But most importantly, the first thing we did when we had the A7V266 crashes was to try the enterprise kernel from LM8.1. And it crashed just as easily as the LM8.2 kernel. That is why I think that we have two problems. These two PCs belong to my colleagues who were very frustrated. Now they have found a solution by using the enterprise kernel from LM8.1 running in LM8.2 they are not terribly interseste in pursuing this problem any further. I think that the problem should really be sorted out. Are there one or two problems (as I think there are)? By making a google search I found that Thomas Wanka had a highmem smp freeze problem with a Serverworks chipset. Marcelo Tosatti answers this question: The ServerWorks chipsets are problematic. Use the noapic boot option. Is this really a third problem. Was it solved by the noapic option? Or do we really have a problem with highmem and 2.4.18?? -- Bjarne On Wed, 2002-05-15 at 19:36, Nicolas ROBAUX wrote: Le Mercredi 15 Mai 2002 19:10, vous avez å¶?it : I need some help (again): We have encountered a total freeze while copying files from an SGI/IRIX system (using NFS) onto a local disk on a PC with an Elitegroup K7S6A/SiS745 MoBo, and more than 1Gb RAM, running the LM 8.2 enterprise kernel 2.4.18-6mdkenterprise. This was repeated many times on two different PCs having identical motherboards (K7S6A/SiS745). Then the same experiment was done with the standard LM8.2 2.4.18-6mdk kernel. Exactly same thing with a MSI 6330 Lite 3 (VIA KT133A Chipset) / Duron 900, standard kernel of the 8.2, on a linux box running as a gateway (no X serveur running, pure console) from cable-modem (Motorola Docsis Surfboard) to wireless personnal network (gatewaydev is Elsa PCI-11). Freezes randomly 2 or 3 times per day, but seems to happen more often on network activity N. Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com -- Nicolas ROBAUX, Journaliste - rédacteur SVM, 2-6, rue des Bourets, BP 145 92154 SURESNES CEDEX FRANCE Tél : +33 (0)1 41 44 40 91 ; Fax : +33 (0)1 41 44 40 90 Pour se rendre à SVM : http://www.mappy.com/PlanPerso/1014683/2 --- Pour des documents sans virus et universels : merci d'envoyer vos textes au format .RTF, et pas .DOC --- Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] K7S6A/SiS745+2.4.18+HIGHMEM freeze(news)
Don't know enough on this to say anything with authority, but I am curious about one thing. Near as I can tell the stock kernel with 8.1 came out before the change in the LVM and the ensuing wars the version in 8.2 has the new one. Could this be the cause of the trouble? I'm grabbing at straws I know but it is a possibility, however slight you may feel, that the problem isn't Mandrake but rather the kernel itself. Has anyone tried an un-Mandraked kernel to see if the same conditions appear? James On Fri, 17 May 2002 11:05:50 +0200 Nicolas ROBAUX [EMAIL PROTECTED] wrote: Same thing for me : I insert noapic in GRUB, but my gateway still hangs randomly. I'll try to recompile the 8.2 kernel, and/or to downgrade to the 8.1 kernel (hope that the 8.1 kernel will run fine with the 8.2 distribution). N. Le Vendredi 17 Mai 2002 04:14, vous avez écrit : hi: noapic didn't work for me. the server will hang under high loading after 4 or 5 hours. i use ctcs testing program to test. mdk 8.1 or red-hat 7.2 didn't have this problem. sorry i didn't have a chance to test it with red-hat 7.3, bc4 the server is one line with mdk 8.1 now... I should report that I got a message from Tom Wanka. The noapic boot option worked for him (he had a serverworks chipset), so this will probably also work for you. We do not use this chipset, so it is not clear if it will work in our case. There could also be a bug in the kernel. I found this interesting line in Alan Cox's description of changes in the 2.4.19 kernel: o Fix highmem truncation on DMA mapping bug (Dave Miller) I do not know if this bug was introduced after 2.4.18, or if it is in 2.4.18. Bjarne On Thu, 2002-05-16 at 08:48, [EMAIL PROTECTED] wrote: hi: i have same problem. 2.4.18 with serverworks and himem, the system hang. i report this several times. but no support or reply yet. Have you seen this! [RHBA-2002:085-11] Kernel panic on SMP systems with ext3 file systems is now fixed. I am not sure if this problem is solved in the Mandrake kernel. -- Bjarne On Wed, 2002-05-15 at 21:58, Bjarne Thomsen wrote: Actually it is on purpose that I did not mention VIA and the PCI latency problem in the subject line. I believe that we have two different problems (unfortunately). The K7S6A was a replacement for an Asus A7V266E with a VIA KT266A chipset. The reson was that we had similar lock-ups, often accompanied by massive disk corruptions which we have not seen with the K7S6A. But most importantly, the first thing we did when we had the A7V266 crashes was to try the enterprise kernel from LM8.1. And it crashed just as easily as the LM8.2 kernel. That is why I think that we have two problems. These two PCs belong to my colleagues who were very frustrated. Now they have found a solution by using the enterprise kernel from LM8.1 running in LM8.2 they are not terribly interseste in pursuing this problem any further. I think that the problem should really be sorted out. Are there one or two problems (as I think there are)? By making a google search I found that Thomas Wanka had a highmem smp freeze problem with a Serverworks chipset. Marcelo Tosatti answers this question: The ServerWorks chipsets are problematic. Use the noapic boot option. Is this really a third problem. Was it solved by the noapic option? Or do we really have a problem with highmem and 2.4.18?? -- Bjarne On Wed, 2002-05-15 at 19:36, Nicolas ROBAUX wrote: Le Mercredi 15 Mai 2002 19:10, vous avez å¶?it : I need some help (again): We have encountered a total freeze while copying files from an SGI/IRIX system (using NFS) onto a local disk on a PC with an Elitegroup K7S6A/SiS745 MoBo, and more than 1Gb RAM, running the LM 8.2 enterprise kernel 2.4.18-6mdkenterprise. This was repeated many times on two different PCs having identical motherboards (K7S6A/SiS745). Then the same experiment was done with the standard LM8.2 2.4.18-6mdk kernel. Exactly same thing with a MSI 6330 Lite 3 (VIA KT133A Chipset) / Duron 900, standard kernel of the 8.2, on a linux box running as a gateway (no X serveur running, pure console) from cable-modem (Motorola Docsis Surfboard) to wireless personnal network (gatewaydev is Elsa PCI-11). Freezes randomly 2 or 3 times per day, but seems to happen more often on network activity N. Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com Want to buy your Pack or Services from MandrakeSoft? Go to
Re: [expert] K7S6A/SiS745+2.4.18+HIGHMEM freeze(news)
hi: i have same problem. 2.4.18 with serverworks and himem, the system hang. i report this several times. but no support or reply yet. Have you seen this! [RHBA-2002:085-11] Kernel panic on SMP systems with ext3 file systems is now fixed. I am not sure if this problem is solved in the Mandrake kernel. -- Bjarne On Wed, 2002-05-15 at 21:58, Bjarne Thomsen wrote: Actually it is on purpose that I did not mention VIA and the PCI latency problem in the subject line. I believe that we have two different problems (unfortunately). The K7S6A was a replacement for an Asus A7V266E with a VIA KT266A chipset. The reson was that we had similar lock-ups, often accompanied by massive disk corruptions which we have not seen with the K7S6A. But most importantly, the first thing we did when we had the A7V266 crashes was to try the enterprise kernel from LM8.1. And it crashed just as easily as the LM8.2 kernel. That is why I think that we have two problems. These two PCs belong to my colleagues who were very frustrated. Now they have found a solution by using the enterprise kernel from LM8.1 running in LM8.2 they are not terribly interseste in pursuing this problem any further. I think that the problem should really be sorted out. Are there one or two problems (as I think there are)? By making a google search I found that Thomas Wanka had a highmem smp freeze problem with a Serverworks chipset. Marcelo Tosatti answers this question: The ServerWorks chipsets are problematic. Use the noapic boot option. Is this really a third problem. Was it solved by the noapic option? Or do we really have a problem with highmem and 2.4.18?? -- Bjarne On Wed, 2002-05-15 at 19:36, Nicolas ROBAUX wrote: Le Mercredi 15 Mai 2002 19:10, vous avez écrit : I need some help (again): We have encountered a total freeze while copying files from an SGI/IRIX system (using NFS) onto a local disk on a PC with an Elitegroup K7S6A/SiS745 MoBo, and more than 1Gb RAM, running the LM 8.2 enterprise kernel 2.4.18-6mdkenterprise. This was repeated many times on two different PCs having identical motherboards (K7S6A/SiS745). Then the same experiment was done with the standard LM8.2 2.4.18-6mdk kernel. Exactly same thing with a MSI 6330 Lite 3 (VIA KT133A Chipset) / Duron 900, standard kernel of the 8.2, on a linux box running as a gateway (no X serveur running, pure console) from cable-modem (Motorola Docsis Surfboard) to wireless personnal network (gatewaydev is Elsa PCI-11). Freezes randomly 2 or 3 times per day, but seems to happen more often on network activity N. Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] K7S6A/SiS745+2.4.18+HIGHMEM freeze
Le Mercredi 15 Mai 2002 19:10, vous avez écrit : Would it possibly help to boot the enterprise kernel by the noapic option? Or is there really a problem with HIGHMEM in the 2.4.18 kernel? Do you have a multi-processor system ? Do you think that the noapic options can be used with one-processor systems ? N. Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] K7S6A/SiS745+2.4.18+HIGHMEM freeze(news)
Le Mercredi 15 Mai 2002 22:05, vous avez écrit : Have you seen this! [RHBA-2002:085-11] Kernel panic on SMP systems with ext3 file systems is now fixed. Nope, I didn't see that (I even don't know where it is from ! An HISTORY file ???) But I don't have a SMP system... N. Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] K7S6A/SiS745+2.4.18+HIGHMEM freeze
The enterprise kernel is actually a combination of SMP and HIGHMEN options in order not to have too many standard kernels. I do not need the SMP option only the HIGHMEM option. This is necessary if you have more than about 900 Mb RAM. I know that our freeze problem is caused by the HIGHMEM option. I recompiled the kernel without SMP but with HIGHMEM, and it still froze. Even though a problem is labeled SMP it could actually very well be a HIGHMEM problem. -- Bjarne On Thu, 2002-05-16 at 11:14, Nicolas ROBAUX wrote: Le Mercredi 15 Mai 2002 19:10, vous avez écrit : Would it possibly help to boot the enterprise kernel by the noapic option? Or is there really a problem with HIGHMEM in the 2.4.18 kernel? Do you have a multi-processor system ? Do you think that the noapic options can be used with one-processor systems ? N. Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] K7S6A/SiS745+2.4.18+HIGHMEM freeze(news)
I should report that I got a message from Tom Wanka. The noapic boot option worked for him (he had a serverworks chipset), so this will probably also work for you. We do not use this chipset, so it is not clear if it will work in our case. There could also be a bug in the kernel. I found this interesting line in Alan Cox's description of changes in the 2.4.19 kernel: o Fix highmem truncation on DMA mapping bug (Dave Miller) I do not know if this bug was introduced after 2.4.18, or if it is in 2.4.18. Bjarne On Thu, 2002-05-16 at 08:48, [EMAIL PROTECTED] wrote: hi: i have same problem. 2.4.18 with serverworks and himem, the system hang. i report this several times. but no support or reply yet. Have you seen this! [RHBA-2002:085-11] Kernel panic on SMP systems with ext3 file systems is now fixed. I am not sure if this problem is solved in the Mandrake kernel. -- Bjarne On Wed, 2002-05-15 at 21:58, Bjarne Thomsen wrote: Actually it is on purpose that I did not mention VIA and the PCI latency problem in the subject line. I believe that we have two different problems (unfortunately). The K7S6A was a replacement for an Asus A7V266E with a VIA KT266A chipset. The reson was that we had similar lock-ups, often accompanied by massive disk corruptions which we have not seen with the K7S6A. But most importantly, the first thing we did when we had the A7V266 crashes was to try the enterprise kernel from LM8.1. And it crashed just as easily as the LM8.2 kernel. That is why I think that we have two problems. These two PCs belong to my colleagues who were very frustrated. Now they have found a solution by using the enterprise kernel from LM8.1 running in LM8.2 they are not terribly interseste in pursuing this problem any further. I think that the problem should really be sorted out. Are there one or two problems (as I think there are)? By making a google search I found that Thomas Wanka had a highmem smp freeze problem with a Serverworks chipset. Marcelo Tosatti answers this question: The ServerWorks chipsets are problematic. Use the noapic boot option. Is this really a third problem. Was it solved by the noapic option? Or do we really have a problem with highmem and 2.4.18?? -- Bjarne On Wed, 2002-05-15 at 19:36, Nicolas ROBAUX wrote: Le Mercredi 15 Mai 2002 19:10, vous avez 嶰rit : I need some help (again): We have encountered a total freeze while copying files from an SGI/IRIX system (using NFS) onto a local disk on a PC with an Elitegroup K7S6A/SiS745 MoBo, and more than 1Gb RAM, running the LM 8.2 enterprise kernel 2.4.18-6mdkenterprise. This was repeated many times on two different PCs having identical motherboards (K7S6A/SiS745). Then the same experiment was done with the standard LM8.2 2.4.18-6mdk kernel. Exactly same thing with a MSI 6330 Lite 3 (VIA KT133A Chipset) / Duron 900, standard kernel of the 8.2, on a linux box running as a gateway (no X serveur running, pure console) from cable-modem (Motorola Docsis Surfboard) to wireless personnal network (gatewaydev is Elsa PCI-11). Freezes randomly 2 or 3 times per day, but seems to happen more often on network activity N. Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] K7S6A/SiS745+2.4.18+HIGHMEM freeze(news)
hi: noapic didn't work for me. the server will hang under high loading after 4 or 5 hours. i use ctcs testing program to test. mdk 8.1 or red-hat 7.2 didn't have this problem. sorry i didn't have a chance to test it with red-hat 7.3, bc4 the server is one line with mdk 8.1 now... I should report that I got a message from Tom Wanka. The noapic boot option worked for him (he had a serverworks chipset), so this will probably also work for you. We do not use this chipset, so it is not clear if it will work in our case. There could also be a bug in the kernel. I found this interesting line in Alan Cox's description of changes in the 2.4.19 kernel: o Fix highmem truncation on DMA mapping bug (Dave Miller) I do not know if this bug was introduced after 2.4.18, or if it is in 2.4.18. Bjarne On Thu, 2002-05-16 at 08:48, [EMAIL PROTECTED] wrote: hi: i have same problem. 2.4.18 with serverworks and himem, the system hang. i report this several times. but no support or reply yet. Have you seen this! [RHBA-2002:085-11] Kernel panic on SMP systems with ext3 file systems is now fixed. I am not sure if this problem is solved in the Mandrake kernel. -- Bjarne On Wed, 2002-05-15 at 21:58, Bjarne Thomsen wrote: Actually it is on purpose that I did not mention VIA and the PCI latency problem in the subject line. I believe that we have two different problems (unfortunately). The K7S6A was a replacement for an Asus A7V266E with a VIA KT266A chipset. The reson was that we had similar lock-ups, often accompanied by massive disk corruptions which we have not seen with the K7S6A. But most importantly, the first thing we did when we had the A7V266 crashes was to try the enterprise kernel from LM8.1. And it crashed just as easily as the LM8.2 kernel. That is why I think that we have two problems. These two PCs belong to my colleagues who were very frustrated. Now they have found a solution by using the enterprise kernel from LM8.1 running in LM8.2 they are not terribly interseste in pursuing this problem any further. I think that the problem should really be sorted out. Are there one or two problems (as I think there are)? By making a google search I found that Thomas Wanka had a highmem smp freeze problem with a Serverworks chipset. Marcelo Tosatti answers this question: The ServerWorks chipsets are problematic. Use the noapic boot option. Is this really a third problem. Was it solved by the noapic option? Or do we really have a problem with highmem and 2.4.18?? -- Bjarne On Wed, 2002-05-15 at 19:36, Nicolas ROBAUX wrote: Le Mercredi 15 Mai 2002 19:10, vous avez 嶰rit : I need some help (again): We have encountered a total freeze while copying files from an SGI/IRIX system (using NFS) onto a local disk on a PC with an Elitegroup K7S6A/SiS745 MoBo, and more than 1Gb RAM, running the LM 8.2 enterprise kernel 2.4.18-6mdkenterprise. This was repeated many times on two different PCs having identical motherboards (K7S6A/SiS745). Then the same experiment was done with the standard LM8.2 2.4.18-6mdk kernel. Exactly same thing with a MSI 6330 Lite 3 (VIA KT133A Chipset) / Duron 900, standard kernel of the 8.2, on a linux box running as a gateway (no X serveur running, pure console) from cable-modem (Motorola Docsis Surfboard) to wireless personnal network (gatewaydev is Elsa PCI-11). Freezes randomly 2 or 3 times per day, but seems to happen more often on network activity N. Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] K7S6A/SiS745+2.4.18+HIGHMEM freeze
Le Mercredi 15 Mai 2002 19:10, vous avez écrit : I need some help (again): We have encountered a total freeze while copying files from an SGI/IRIX system (using NFS) onto a local disk on a PC with an Elitegroup K7S6A/SiS745 MoBo, and more than 1Gb RAM, running the LM 8.2 enterprise kernel 2.4.18-6mdkenterprise. This was repeated many times on two different PCs having identical motherboards (K7S6A/SiS745). Then the same experiment was done with the standard LM8.2 2.4.18-6mdk kernel. Exactly same thing with a MSI 6330 Lite 3 (VIA KT133A Chipset) / Duron 900, standard kernel of the 8.2, on a linux box running as a gateway (no X serveur running, pure console) from cable-modem (Motorola Docsis Surfboard) to wireless personnal network (gatewaydev is Elsa PCI-11). Freezes randomly 2 or 3 times per day, but seems to happen more often on network activity N. Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] K7S6A/SiS745+2.4.18+HIGHMEM freeze
Actually it is on purpose that I did not mention VIA and the PCI latency problem in the subject line. I believe that we have two different problems (unfortunately). The K7S6A was a replacement for an Asus A7V266E with a VIA KT266A chipset. The reson was that we had similar lock-ups, often accompanied by massive disk corruptions which we have not seen with the K7S6A. But most importantly, the first thing we did when we had the A7V266 crashes was to try the enterprise kernel from LM8.1. And it crashed just as easily as the LM8.2 kernel. That is why I think that we have two problems. These two PCs belong to my colleagues who were very frustrated. Now they have found a solution by using the enterprise kernel from LM8.1 running in LM8.2 they are not terribly interseste in pursuing this problem any further. I think that the problem should really be sorted out. Are there one or two problems (as I think there are)? By making a google search I found that Thomas Wanka had a highmem smp freeze problem with a Serverworks chipset. Marcelo Tosatti answers this question: The ServerWorks chipsets are problematic. Use the noapic boot option. Is this really a third problem. Was it solved by the noapic option? Or do we really have a problem with highmem and 2.4.18?? -- Bjarne On Wed, 2002-05-15 at 19:36, Nicolas ROBAUX wrote: Le Mercredi 15 Mai 2002 19:10, vous avez écrit : I need some help (again): We have encountered a total freeze while copying files from an SGI/IRIX system (using NFS) onto a local disk on a PC with an Elitegroup K7S6A/SiS745 MoBo, and more than 1Gb RAM, running the LM 8.2 enterprise kernel 2.4.18-6mdkenterprise. This was repeated many times on two different PCs having identical motherboards (K7S6A/SiS745). Then the same experiment was done with the standard LM8.2 2.4.18-6mdk kernel. Exactly same thing with a MSI 6330 Lite 3 (VIA KT133A Chipset) / Duron 900, standard kernel of the 8.2, on a linux box running as a gateway (no X serveur running, pure console) from cable-modem (Motorola Docsis Surfboard) to wireless personnal network (gatewaydev is Elsa PCI-11). Freezes randomly 2 or 3 times per day, but seems to happen more often on network activity N. Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] K7S6A/SiS745+2.4.18+HIGHMEM freeze(news)
Have you seen this! [RHBA-2002:085-11] Kernel panic on SMP systems with ext3 file systems is now fixed. I am not sure if this problem is solved in the Mandrake kernel. -- Bjarne On Wed, 2002-05-15 at 21:58, Bjarne Thomsen wrote: Actually it is on purpose that I did not mention VIA and the PCI latency problem in the subject line. I believe that we have two different problems (unfortunately). The K7S6A was a replacement for an Asus A7V266E with a VIA KT266A chipset. The reson was that we had similar lock-ups, often accompanied by massive disk corruptions which we have not seen with the K7S6A. But most importantly, the first thing we did when we had the A7V266 crashes was to try the enterprise kernel from LM8.1. And it crashed just as easily as the LM8.2 kernel. That is why I think that we have two problems. These two PCs belong to my colleagues who were very frustrated. Now they have found a solution by using the enterprise kernel from LM8.1 running in LM8.2 they are not terribly interseste in pursuing this problem any further. I think that the problem should really be sorted out. Are there one or two problems (as I think there are)? By making a google search I found that Thomas Wanka had a highmem smp freeze problem with a Serverworks chipset. Marcelo Tosatti answers this question: The ServerWorks chipsets are problematic. Use the noapic boot option. Is this really a third problem. Was it solved by the noapic option? Or do we really have a problem with highmem and 2.4.18?? -- Bjarne On Wed, 2002-05-15 at 19:36, Nicolas ROBAUX wrote: Le Mercredi 15 Mai 2002 19:10, vous avez écrit : I need some help (again): We have encountered a total freeze while copying files from an SGI/IRIX system (using NFS) onto a local disk on a PC with an Elitegroup K7S6A/SiS745 MoBo, and more than 1Gb RAM, running the LM 8.2 enterprise kernel 2.4.18-6mdkenterprise. This was repeated many times on two different PCs having identical motherboards (K7S6A/SiS745). Then the same experiment was done with the standard LM8.2 2.4.18-6mdk kernel. Exactly same thing with a MSI 6330 Lite 3 (VIA KT133A Chipset) / Duron 900, standard kernel of the 8.2, on a linux box running as a gateway (no X serveur running, pure console) from cable-modem (Motorola Docsis Surfboard) to wireless personnal network (gatewaydev is Elsa PCI-11). Freezes randomly 2 or 3 times per day, but seems to happen more often on network activity N. Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com