Re: Nvidia Upgrade Agony
Paulo Cavalcanti <[email protected]> Mon, 5 Jan 2015 05:52:08 -0200
| Newsgroups | gmane.linux.redhat.rpm.atrpms.general |
|---|---|
| Message-ID | <CAMgUmn==8Eq2a_1c1wQau6UfUoDD9mFUN+GwLF4s3jY3c73viQ@mail.gmail.com> |
--===============3128583314901659782== Content-Type: multipart/alternative; boundary=001a113a7c029aefd3050be2f7f3 --001a113a7c029aefd3050be2f7f3 Content-Type: text/plain; charset=UTF-8 Hi, I need a set of nvidia rpms for an old rhel6 dell notebook with a nvidia 8400. Since I upgrades to rhel 6.6, the ones I used stopped to work. Does anyone know what is the appropriate version now? Also, I have a new le novo notebook running f21. Anyone has experience with bumblebee and nvidia optimus technology? Definitively, nouveau did not work with bumblebee ... By the way, atrpms-rpm-config need to be fixed for f21. The place macros are kept now has changed, and the original package is creating rpm names with f21 instead of fc21. For now, I created a symbolic link for macros.dist to /etc/rpm. Thanks. On Mon, Dec 29, 2014 at 4:10 AM, Stephen Collier <[email protected]> wrote: > On 29/12/2014 4:30 pm, Stephen Collier wrote: > >> On 29/12/2014 8:08 am, Larry Kennedy wrote: >> >>> I'm running myth 0.26.1-296 on EL6 (2.6.32-431.el6.i686) with a GT210 >>> card and the nvidia 319.32 driver. >>> >>> About a week ago, I accidentally took an update to my Xorg server (now I >>> have xorg-x11-server-Xorg-1.15.0-25.el6.centos.i686), and after that I >>> started getting a driver mismatch error (ABI 15 I seem to recall). Anyway, >>> I've seen this before, and I believe this means the nvidia 319 driver is at >>> a lower version than is needed, so I figured all I need is an nvidia >>> upgrade. If I remember right, the driver from atrpms was still far enough >>> behind to cause the ABI 15 error, so I tried kmod-nvidia from elrepo. >>> After removing the old nvidia 319 driver, this update pulled in these >>> nvidia libraries: >>> >>> nvidia-x11-drv-340.65-1.el6.elrepo.i686 >>> kmod-nvidia-340.65-1.el6.elrepo.i686 >>> >>> There is no nvidia kmdl rpm installed now (as was the case before), so I >>> am not sure if this is a problem or not. When I start the desktop, >>> everything seems fine, and apps like Mozilla work, but none of the myth >>> GUIs will start. I get a segmentation fault with both the mythfrontend and >>> mythtv-setup. Even with --verbose, mythfrontend gives me nothing to work >>> with. >>> >>> As a test, I also tried doing the nvidia upgrade using scrpms. With >>> this upgrade, I was forced to move up to kernel-2.6.32-504.3.3.el6.i686, >>> so at least I know this update is pulling in an nvidia-kmdl, unlike kmod. >>> In the end, this upgrade attempt was less successful than the kmod upgrade >>> in that the nvidia driver is never able to load, so the desktop never >>> starts. At the bottom of Xorg.0.log it shows this: >>> >>> [ 20.723] (II) LoadModule: "nvidia" >>> [ 20.723] (WW) Warning, couldn't open module nvidia >>> [ 20.723] (II) UnloadModule: "nvidia" >>> [ 20.723] (II) Unloading nvidia >>> [ 20.723] (EE) Failed to load module "nvidia" (module does not exist, >>> 0) >>> [ 20.723] (EE) No drivers available. >>> [ 20.723] (EE) Fatal server error: >>> [ 20.723] (EE) no screens found(EE) >>> [ 20.723] (EE) Please consult the CentOS support at >>> http://wiki.centos.org/Documentation >>> >>> This seems odd since "modprobe nvidia" succeeds. I tried nvidia-xconfig >>> to recreate the xorg.conf, but no go. Of course I am rebooting every time >>> to get a clean start. After many attempts, it looks like kmod-nvidia >>> 340.65 may be my best option. >>> >>> Is anyone aware of an issue with mythtv 0.26 and the nvidia 340.xx >>> driver? Alternatively, if anyone has ideas on how to get the scrpms nvidia >>> driver to work, I am open to trying that again too. >>> >>> I have options, but nothing looks good. Downgrading Xorg seems >>> problematic, so I didn't go there. I suppose I could upgrade myth to >>> version 0.27 and see what happens, but that presents some difficulty on EL6 >>> (although I know that scrpms has compiled both myth 0.27 and the compatible >>> version of Qt). It also triggers multiple myth upgrades for BE-FE >>> compatibility. Another option is to go to EL7, but that requires a decent >>> effort too since there is no good upgrade-in-place from EL6 32-bit to EL7). >>> >>> I apologize in advance if this post is somewhat schizophrenic. I can >>> provide more details if anyone can lend a hand. >>> >>> Larry >>> >>> I'm sorry for the delay - I'll have a look and do some testing with my >> el6 frontend I also upgraded a from el6 32bit to 64 as a path to el7 If you >> want I'll send you the process. >> >> Cheers >> Stephen >> >> _______________________________________________ >> atrpms-users mailing list >> [email protected] >> http://lists.atrpms.net/mailman/listinfo/atrpms-users >> >> I've found the problem and fixed it. The src rpms are built - the > updated nvidia drivers will build over the next hour or so. > > > Cheers > Stephen > > _______________________________________________ > atrpms-users mailing list > [email protected] > http://lists.atrpms.net/mailman/listinfo/atrpms-users > -- Paulo Roma Cavalcanti LCG - UFRJ --001a113a7c029aefd3050be2f7f3 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable <div dir=3D"ltr"><div><div><div><div><div><div>Hi,<br><br></div>I need a se= t of nvidia rpms for an old rhel6 dell notebook with a nvidia 8400.<br><br>= </div>Since I upgrades to rhel 6.6, the ones I used stopped to work.<br><br= ></div>Does anyone know what is the appropriate version now?<br><br></div>A= lso, I have a new le novo notebook running f21. Anyone has experience with = bumblebee and nvidia optimus technology?<br></div>Definitively, nouveau did= not work with bumblebee ...<br><br></div><div>By the way, atrpms-rpm-confi= g need to be fixed for f21. The place macros are kept now has changed,<br><= /div><div>and the original package is creating rpm names with f21 instead o= f fc21. For now, I created a symbolic<br></div><div>link for macros.dist to= /etc/rpm. <br></div><div><br></div>Thanks. <br><div><div><div><br>=C2=A0 <= br></div></div></div></div><div class=3D"gmail_extra"><br><div class=3D"gma= il_quote">On Mon, Dec 29, 2014 at 4:10 AM, Stephen Collier <span dir=3D"ltr= "><<a href=3D"mailto:[email protected]" target=3D"_blank">judithc@b= igpond.net.au</a>></span> wrote:<br><blockquote class=3D"gmail_quote" st= yle=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div = class=3D"HOEnZb"><div class=3D"h5">On 29/12/2014 4:30 pm, Stephen Collier w= rote:<br> <blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1p= x #ccc solid;padding-left:1ex"> On 29/12/2014 8:08 am, Larry Kennedy wrote:<br> <blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1p= x #ccc solid;padding-left:1ex"> I'm running myth 0.26.1-296 on EL6 (2.6.32-431.el6.i686) with a GT210 c= ard and the nvidia 319.32 driver.<br> <br> About a week ago, I accidentally took an update to my Xorg server (now I ha= ve xorg-x11-server-Xorg-1.15.0-<u></u>25.el6.centos.i686), and after that I= started getting a driver mismatch error (ABI 15 I seem to recall).=C2=A0 A= nyway, I've seen this before, and I believe this means the nvidia 319 d= river is at a lower version than is needed, so I figured all I need is an n= vidia upgrade. If I remember right, the driver from atrpms was still far en= ough behind to cause the ABI 15 error, so I tried kmod-nvidia from elrepo.= =C2=A0 After removing the old nvidia 319 driver, this update pulled in thes= e nvidia libraries:<br> <br> nvidia-x11-drv-340.65-1.el6.<u></u>elrepo.i686<br> kmod-nvidia-340.65-1.el6.<u></u>elrepo.i686<br> <br> There is no nvidia kmdl rpm installed now (as was the case before), so I am= not sure if this is a problem or not.=C2=A0 When I start the desktop, ever= ything seems fine, and apps like Mozilla work, but none of the myth GUIs wi= ll start.=C2=A0 I get a segmentation fault with both the mythfrontend and m= ythtv-setup. Even with --verbose, mythfrontend gives me nothing to work wit= h.<br> <br> As a test, I also tried doing the nvidia upgrade using scrpms.=C2=A0 With t= his upgrade, I was forced to move up to kernel-2.6.32-504.3.3.el6.<u></u>i6= 86, so at least I know this update is pulling in an nvidia-kmdl, unlike kmo= d.=C2=A0 In the end, this upgrade attempt was less successful than the kmod= upgrade in that the nvidia driver is never able to load, so the desktop ne= ver starts. At the bottom of Xorg.0.log it shows this:<br> <br> [=C2=A0 =C2=A0 20.723] (II) LoadModule: "nvidia"<br> [=C2=A0 =C2=A0 20.723] (WW) Warning, couldn't open module nvidia<br> [=C2=A0 =C2=A0 20.723] (II) UnloadModule: "nvidia"<br> [=C2=A0 =C2=A0 20.723] (II) Unloading nvidia<br> [=C2=A0 =C2=A0 20.723] (EE) Failed to load module "nvidia" (modul= e does not exist, 0)<br> [=C2=A0 =C2=A0 20.723] (EE) No drivers available.<br> [=C2=A0 =C2=A0 20.723] (EE) Fatal server error:<br> [=C2=A0 =C2=A0 20.723] (EE) no screens found(EE)<br> [=C2=A0 =C2=A0 20.723] (EE) Please consult the CentOS support at <a href=3D= "http://wiki.centos.org/Documentation" target=3D"_blank">http://wiki.centos= .org/<u></u>Documentation</a><br> <br> This seems odd since "modprobe nvidia" succeeds.=C2=A0 I tried nv= idia-xconfig to recreate the xorg.conf, but no go. Of course I am rebooting= every time to get a clean start.=C2=A0 After many attempts, it looks like = kmod-nvidia 340.65 may be my best option.<br> <br> Is anyone aware of an issue with mythtv 0.26 and the nvidia 340.xx driver?= =C2=A0 Alternatively, if anyone has ideas on how to get the scrpms nvidia d= river to work, I am open to trying that again too.<br> <br> I have options, but nothing looks good.=C2=A0 Downgrading Xorg seems proble= matic, so I didn't go there. I suppose I could upgrade myth to version = 0.27 and see what happens, but that presents some difficulty on EL6 (althou= gh I know that scrpms has compiled both myth 0.27 and the compatible versio= n of Qt). It also triggers multiple myth upgrades for BE-FE compatibility. = Another option is to go to EL7, but that requires a decent effort too since= there is no good upgrade-in-place from EL6 32-bit to EL7).<br> <br> I apologize in advance if this post is somewhat schizophrenic.=C2=A0 I can = provide more details if anyone can lend a hand.<br> <br> Larry<br> <br> </blockquote> I'm sorry for the delay - I'll have a look and do some testing with= my el6 frontend I also upgraded a from el6 32bit to 64 as a path to el7 If= you want I'll send you the process.<br> <br> Cheers<br> Stephen<br> <br> ______________________________<u></u>_________________<br> atrpms-users mailing list<br> <a href=3D"mailto:[email protected]" target=3D"_blank">atrpms-users@a= trpms.net</a><br> <a href=3D"http://lists.atrpms.net/mailman/listinfo/atrpms-users" target=3D= "_blank">http://lists.atrpms.net/<u></u>mailman/listinfo/atrpms-users</a><b= r> <br> </blockquote></div></div> I've found the problem and fixed it. The src rpms are built - the updat= ed nvidia drivers will build over the next hour or so.<div class=3D"HOEnZb"= ><div class=3D"h5"><br> <br> Cheers<br> Stephen<br> <br> ______________________________<u></u>_________________<br> atrpms-users mailing list<br> <a href=3D"mailto:[email protected]" target=3D"_blank">atrpms-users@a= trpms.net</a><br> <a href=3D"http://lists.atrpms.net/mailman/listinfo/atrpms-users" target=3D= "_blank">http://lists.atrpms.net/<u></u>mailman/listinfo/atrpms-users</a><b= r> </div></div></blockquote></div><br><br clear=3D"all"><br>-- <br><div class= =3D"gmail_signature"><div dir=3D"ltr">Paulo Roma Cavalcanti<br>LCG - UFRJ</= div></div> </div> --001a113a7c029aefd3050be2f7f3-- --===============3128583314901659782== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ atrpms-users mailing list [email protected] http://lists.atrpms.net/mailman/listinfo/atrpms-users --===============3128583314901659782==--