From dsavage at peaknet.net Thu Feb 2 12:19:55 2012 From: dsavage at peaknet.net (Robert G. (Doc) Savage) Date: Thu, 02 Feb 2012 06:19:55 -0600 Subject: [rhelv6-list] rtkit-daemon looping leads to crash Message-ID: <1328185195.3520.16.camel@lion.protogeek.org> Following a complete re-install, I found (IIRC) the only real user on my RHEL6 box had uid/gui of 501/502. Somehow the xguest user got assigned its uid/gid ahead of the real user. To match the underlying uid/gid assignments for that real user on surrounding machines, I removed the xguest package and reassigned 501:501 to the real user. Now I'm seeing some serious thrashing of the rtkit-daemon process that causes the machine to crash in 2-3 hours. The following lines repeat in /var/log/messages with a new rtkit-daemon PID every few minutes. Can someone suggest where to start untying this knot? Feb 2 05:58:42 lion rtkit-daemon[3061]: Sucessfully made thread 3239 of process 3239 (/usr/bin/pulseaudio) owned by '501' high priority at nice level -11. Feb 2 05:58:42 lion pulseaudio[3239]: pid.c: Stale PID file, overwriting. Feb 2 05:58:42 lion rtkit-daemon[3061]: Sucessfully made thread 3241 of process 3239 (/usr/bin/pulseaudio) owned by '501' RT at priority 5. Feb 2 05:58:42 lion rtkit-daemon[3061]: Sucessfully made thread 3242 of process 3239 (/usr/bin/pulseaudio) owned by '501' RT at priority 5. Feb 2 05:58:44 lion rtkit-daemon[3061]: Sucessfully made thread 3349 of process 3349 (/usr/bin/pulseaudio) owned by '501' high priority at nice level -11. Feb 2 05:58:44 lion pulseaudio[3349]: pid.c: Daemon already running. --Doc Savage Fairview Heights, IL From michael.coffman at avagotech.com Thu Feb 9 18:13:24 2012 From: michael.coffman at avagotech.com (Michael Coffman) Date: Thu, 9 Feb 2012 11:13:24 -0700 Subject: [rhelv6-list] Help with kickstart %pre Message-ID: Hello, Is there a way to change the screen resolution that is used during the kickstart %pre section? I use the pre section of the kickstart on RHEL4 and 5 to query the user doing the install for input. RHEL4 and 5 default to vga with fairly large fonts. As of RHEL6 it seems to discover the hardware and then adjusts the display. The result is a very tiny font. I do installs via a pxeboot server and have tried various options to the append line for the pxe kernel with no affect. Any input would be greatly appreciated Thanks. -- -MichaelC From hbrown at divms.uiowa.edu Thu Feb 9 18:20:02 2012 From: hbrown at divms.uiowa.edu (Hugh Brown) Date: Thu, 09 Feb 2012 12:20:02 -0600 Subject: [rhelv6-list] Help with kickstart %pre In-Reply-To: References: Message-ID: <4F340E52.1000806@divms.uiowa.edu> On 02/09/2012 12:13 PM, Michael Coffman wrote: > Hello, > > Is there a way to change the screen resolution that is used during the > kickstart %pre section? I use the pre section of the kickstart on > RHEL4 and 5 to query the user doing the install for input. RHEL4 and > 5 default to vga with fairly large fonts. As of RHEL6 it seems to > discover the hardware and then adjusts the display. The result is a > very tiny font. I do installs via a pxeboot server and have tried > various options to the append line > for the pxe kernel with no affect. Any input would be greatly appreciated > > Thanks. If the machines are using the nouveau driver, you can probably append nouveau.modeset=0 to the pxe command line for the kernel. I think that should give you back the old behavior. You can also put a sleep in the pre section, switch to the vt on F2. cat /proc/cmdline to see how the kernel got started and check to see if there's an obvious parameter that can be adjusted. Hugh From gsgatlin at ncsu.edu Thu Feb 9 18:43:07 2012 From: gsgatlin at ncsu.edu (Gary Gatling) Date: Thu, 9 Feb 2012 13:43:07 -0500 Subject: [rhelv6-list] Help with kickstart %pre In-Reply-To: References: Message-ID: Maybe what will help is to append vga=something to the kernel boot line when you boot the install media. Such as: vga=785 Hope that helps, On Thu, Feb 9, 2012 at 1:13 PM, Michael Coffman < michael.coffman at avagotech.com> wrote: > Hello, > > Is there a way to change the screen resolution that is used during the > kickstart %pre section? I use the pre section of the kickstart on > RHEL4 and 5 to query the user doing the install for input. RHEL4 and > 5 default to vga with fairly large fonts. As of RHEL6 it seems to > discover the hardware and then adjusts the display. The result is a > very tiny font. I do installs via a pxeboot server and have tried > various options to the append line > for the pxe kernel with no affect. Any input would be greatly appreciated > > Thanks. > -- > -MichaelC > > _______________________________________________ > rhelv6-list mailing list > rhelv6-list at redhat.com > https://www.redhat.com/mailman/listinfo/rhelv6-list > -------------- next part -------------- An HTML attachment was scrubbed... URL: From michael.coffman at avagotech.com Thu Feb 9 18:51:02 2012 From: michael.coffman at avagotech.com (Michael Coffman) Date: Thu, 9 Feb 2012 11:51:02 -0700 Subject: [rhelv6-list] Help with kickstart %pre In-Reply-To: <4F340E52.1000806@divms.uiowa.edu> References: <4F340E52.1000806@divms.uiowa.edu> Message-ID: On Thu, Feb 9, 2012 at 11:20 AM, Hugh Brown wrote: > On 02/09/2012 12:13 PM, Michael Coffman wrote: >> >> Hello, >> >> Is there a way to change the screen resolution that is used during the >> kickstart %pre section? ? I use the pre section of the kickstart on >> RHEL4 and 5 to query the user doing the install for input. ? RHEL4 and >> 5 default to vga with fairly large fonts. ? As of RHEL6 it seems to >> discover the hardware and then adjusts the display. ?The result is a >> very tiny font. ? I do installs via a pxeboot server and have tried >> various options to the append line >> for the pxe kernel with no affect. ?Any input would be greatly appreciated >> >> Thanks. > > > If the machines are using the nouveau driver, you can probably append > nouveau.modeset=0 to the pxe command line for the kernel. > > I think that should give you back the old behavior. Thanks much. That worked. > > You can also put a sleep in the pre section, switch to the vt on F2. cat > /proc/cmdline to see how the kernel got started and check to see if there's > an obvious parameter that can be adjusted. > > Hugh -- -MichaelC From michael.coffman at avagotech.com Thu Feb 9 18:51:49 2012 From: michael.coffman at avagotech.com (Michael Coffman) Date: Thu, 9 Feb 2012 11:51:49 -0700 Subject: [rhelv6-list] Help with kickstart %pre In-Reply-To: References: Message-ID: On Thu, Feb 9, 2012 at 11:43 AM, Gary Gatling wrote: > Maybe what will help is to append vga=something to the kernel boot line when > you boot the install media. Such as: > > vga=785 I didn't see any difference seen when changing the vga values. Thanks. > > Hope that helps, > > On Thu, Feb 9, 2012 at 1:13 PM, Michael Coffman > wrote: >> >> Hello, >> >> Is there a way to change the screen resolution that is used during the >> kickstart %pre section? ? I use the pre section of the kickstart on >> RHEL4 and 5 to query the user doing the install for input. ? RHEL4 and >> 5 default to vga with fairly large fonts. ? As of RHEL6 it seems to >> discover the hardware and then adjusts the display. ?The result is a >> very tiny font. ? I do installs via a pxeboot server and have tried >> various options to the append line >> for the pxe kernel with no affect. ?Any input would be greatly appreciated >> >> Thanks. >> -- >> -MichaelC >> >> _______________________________________________ >> rhelv6-list mailing list >> rhelv6-list at redhat.com >> https://www.redhat.com/mailman/listinfo/rhelv6-list > > -- -MichaelC From prentice at ias.edu Thu Feb 9 21:15:13 2012 From: prentice at ias.edu (Prentice Bisbal) Date: Thu, 09 Feb 2012 16:15:13 -0500 Subject: [rhelv6-list] Help with kickstart %pre In-Reply-To: References: <4F340E52.1000806@divms.uiowa.edu> Message-ID: <4F343761.9010904@ias.edu> On 02/09/2012 01:51 PM, Michael Coffman wrote: > On Thu, Feb 9, 2012 at 11:20 AM, Hugh Brown wrote: >> On 02/09/2012 12:13 PM, Michael Coffman wrote: >>> Hello, >>> >>> Is there a way to change the screen resolution that is used during the >>> kickstart %pre section? I use the pre section of the kickstart on >>> RHEL4 and 5 to query the user doing the install for input. RHEL4 and >>> 5 default to vga with fairly large fonts. As of RHEL6 it seems to >>> discover the hardware and then adjusts the display. The result is a >>> very tiny font. I do installs via a pxeboot server and have tried >>> various options to the append line >>> for the pxe kernel with no affect. Any input would be greatly appreciated >>> >>> Thanks. >> >> If the machines are using the nouveau driver, you can probably append >> nouveau.modeset=0 to the pxe command line for the kernel. >> >> I think that should give you back the old behavior. > Thanks much. That worked. > How about just adding 'nomodeset' to your kernel args? That might be more generic/portable. -- Prentice From jjneely at ncsu.edu Thu Feb 9 22:01:18 2012 From: jjneely at ncsu.edu (Jack Neely) Date: Thu, 9 Feb 2012 17:01:18 -0500 Subject: [rhelv6-list] KVM causing interface overruns Message-ID: <20120209220118.GW19880@virge.linuxczar.net> Folks, I had some network performance problems on one of my KVM VMs recently. Switching to the virtio network driver made the VM usable. However, in this process I noticed that the VM host machine has a high number of overruns for the vnetX interfaces attached to each VM. Especially so for UDP-heavy processes such as PXE booting (TFTP). vnet0 Link encap:Ethernet HWaddr FE:52:00:01:01:94 inet6 addr: fe80::fc52:ff:fe01:194/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:1352052 errors:0 dropped:0 overruns:0 frame:0 TX packets:13098674 errors:0 dropped:0 overruns:1527 carrier:0 collisions:0 txqueuelen:500 RX bytes:140258512 (133.7 MiB) TX bytes:959158287 (914.7 MiB) I've seen the overruns much higher before I switched to the virtio network driver. What is causing the overruns? They seem to severely impact network performance. My KVM hosts are RHEL 6.2 using the latest kernel and updates. Jack Neely -- Jack Neely Linux Czar, OIT Campus Linux Services Office of Information Technology, NC State University GPG Fingerprint: 1917 5AC1 E828 9337 7AA4 EA6B 213B 765F 3B6A 5B89 From michael.coffman at avagotech.com Fri Feb 10 15:18:05 2012 From: michael.coffman at avagotech.com (Michael Coffman) Date: Fri, 10 Feb 2012 08:18:05 -0700 Subject: [rhelv6-list] Help with kickstart %pre In-Reply-To: <4F343761.9010904@ias.edu> References: <4F340E52.1000806@divms.uiowa.edu> <4F343761.9010904@ias.edu> Message-ID: On Thu, Feb 9, 2012 at 2:15 PM, Prentice Bisbal wrote: > > On 02/09/2012 01:51 PM, Michael Coffman wrote: >> On Thu, Feb 9, 2012 at 11:20 AM, Hugh Brown wrote: >>> On 02/09/2012 12:13 PM, Michael Coffman wrote: >>>> Hello, >>>> >>>> Is there a way to change the screen resolution that is used during the >>>> kickstart %pre section? ? I use the pre section of the kickstart on >>>> RHEL4 and 5 to query the user doing the install for input. ? RHEL4 and >>>> 5 default to vga with fairly large fonts. ? As of RHEL6 it seems to >>>> discover the hardware and then adjusts the display. ?The result is a >>>> very tiny font. ? I do installs via a pxeboot server and have tried >>>> various options to the append line >>>> for the pxe kernel with no affect. ?Any input would be greatly appreciated >>>> >>>> Thanks. >>> >>> If the machines are using the nouveau driver, you can probably append >>> nouveau.modeset=0 to the pxe command line for the kernel. >>> >>> I think that should give you back the old behavior. >> Thanks much. ?That worked. >> > How about just adding 'nomodeset' to your kernel args? ?That might be > more generic/portable. Even better. Thanks for the pointer. > > -- > Prentice > > -- -MichaelC From akrherz at iastate.edu Mon Feb 13 14:11:10 2012 From: akrherz at iastate.edu (daryl herzmann) Date: Mon, 13 Feb 2012 08:11:10 -0600 (CST) Subject: [rhelv6-list] WARNING: at net/sched/sch_generic.c:261 dev_watchdog leads to unresponsive network Message-ID: Greetings all, I recently (~ 1 Jan) installed RHEL6 64bit on a system that has a: Intel Corporation 82574L Gigabit Network Connection which uses the e1000e module. Anyway, I've had nothing but trouble since. The network device becomes unresponsive after some seemingly random amount of time and a reboot is necessary to get it back functioning (reloading the module does not seem to help). I filed this issue in bugzilla here: https://bugzilla.redhat.com/show_bug.cgi?id=782586 The standard error that comes to the console is: Jan 17 12:38:46 xxx kernel: WARNING: at net/sched/sch_generic.c:261 dev_watchdog+0x26d/0x280() (Not tainted) Jan 17 12:38:46 xxx kernel: Hardware name: X9SCL/X9SCM Jan 17 12:38:46 xxx kernel: NETDEV WATCHDOG: eth0 (e1000e): transmit queu e 0 timed out This system is fully updated RHEL6.2. Anybody seem this before / have ideas on what to try to make it stop! :) thanks, daryl From brilong at cisco.com Mon Feb 13 14:30:51 2012 From: brilong at cisco.com (Brian Long (brilong)) Date: Mon, 13 Feb 2012 14:30:51 +0000 Subject: [rhelv6-list] WARNING: at net/sched/sch_generic.c:261 dev_watchdog leads to unresponsive network In-Reply-To: References: Message-ID: <454D72C6-2514-4C6A-AA14-E693EF2892EB@cisco.com> Could the system have old firmware on that NIC? Does your vendor provide firmware updates which could be installed? Without looking into the problem, I wonder if it's a driver/firmware interaction issue. /Brian/ On Feb 13, 2012, at 9:11 AM, daryl herzmann wrote: > Greetings all, > > I recently (~ 1 Jan) installed RHEL6 64bit on a system that has a: > > Intel Corporation 82574L Gigabit Network Connection > > which uses the e1000e module. Anyway, I've had nothing but trouble since. The network device becomes unresponsive after some seemingly random amount of time and a reboot is necessary to get it back functioning (reloading the module does not seem to help). I filed this issue in bugzilla here: > > https://bugzilla.redhat.com/show_bug.cgi?id=782586 > > The standard error that comes to the console is: > > Jan 17 12:38:46 xxx kernel: WARNING: at net/sched/sch_generic.c:261 dev_watchdog+0x26d/0x280() (Not tainted) > Jan 17 12:38:46 xxx kernel: Hardware name: X9SCL/X9SCM > Jan 17 12:38:46 xxx kernel: NETDEV WATCHDOG: eth0 (e1000e): transmit queu > e 0 timed out > > This system is fully updated RHEL6.2. Anybody seem this before / have ideas on what to try to make it stop! :) > > thanks, > daryl > > _______________________________________________ > rhelv6-list mailing list > rhelv6-list at redhat.com > https://www.redhat.com/mailman/listinfo/rhelv6-list From yossig at emet.co.il Mon Feb 13 16:01:15 2012 From: yossig at emet.co.il (Yossi Goltz) Date: Mon, 13 Feb 2012 18:01:15 +0200 Subject: [rhelv6-list] WARNING: at net/sched/sch_generic.c:261 dev_watchdog leads to unresponsive network In-Reply-To: <454D72C6-2514-4C6A-AA14-E693EF2892EB@cisco.com> Message-ID: <0LZC00JI39U3QBC0@davar.emet.co.il> How about opening a case with Redhat? It might help speed up the process.... /Yossi/ -----Original Message----- From: rhelv6-list-bounces at redhat.com [mailto:rhelv6-list-bounces at redhat.com] On Behalf Of Brian Long (brilong) Sent: Monday, February 13, 2012 4:31 PM To: Red Hat Enterprise Linux 6 (Santiago) discussion mailing-list Subject: Re: [rhelv6-list] WARNING: at net/sched/sch_generic.c:261 dev_watchdog leads to unresponsive network Could the system have old firmware on that NIC? Does your vendor provide firmware updates which could be installed? Without looking into the problem, I wonder if it's a driver/firmware interaction issue. /Brian/ On Feb 13, 2012, at 9:11 AM, daryl herzmann wrote: > Greetings all, > > I recently (~ 1 Jan) installed RHEL6 64bit on a system that has a: > > Intel Corporation 82574L Gigabit Network Connection > > which uses the e1000e module. Anyway, I've had nothing but trouble since. The network device becomes unresponsive after some seemingly random amount of time and a reboot is necessary to get it back functioning (reloading the module does not seem to help). I filed this issue in bugzilla here: > > https://bugzilla.redhat.com/show_bug.cgi?id=782586 > > The standard error that comes to the console is: > > Jan 17 12:38:46 xxx kernel: WARNING: at net/sched/sch_generic.c:261 > dev_watchdog+0x26d/0x280() (Not tainted) Jan 17 12:38:46 xxx kernel: > Hardware name: X9SCL/X9SCM Jan 17 12:38:46 xxx kernel: NETDEV > WATCHDOG: eth0 (e1000e): transmit queu e 0 timed out > > This system is fully updated RHEL6.2. Anybody seem this before / have > ideas on what to try to make it stop! :) > > thanks, > daryl > > _______________________________________________ > rhelv6-list mailing list > rhelv6-list at redhat.com > https://www.redhat.com/mailman/listinfo/rhelv6-list _______________________________________________ rhelv6-list mailing list rhelv6-list at redhat.com https://www.redhat.com/mailman/listinfo/rhelv6-list From rprice at redhat.com Mon Feb 13 16:24:10 2012 From: rprice at redhat.com (Robin Price II) Date: Mon, 13 Feb 2012 11:24:10 -0500 Subject: [rhelv6-list] WARNING: at net/sched/sch_generic.c:261 dev_watchdog leads to unresponsive network In-Reply-To: <0LZC00JI39U3QBC0@davar.emet.co.il> References: <0LZC00JI39U3QBC0@davar.emet.co.il> Message-ID: <4F39392A.7080000@redhat.com> I agree with Yossi here. Bugzilla is not a support tool for Red Hat and I believe you would have better results opening a case. I did find a kbase on this matter for you: kernel: WARNING: at net/sched/sch_generic.c dev_watchdog() on RHEL6 https://access.redhat.com/kb/docs/DOC-48625 Best of luck! ~rp On 02/13/2012 11:01 AM, Yossi Goltz wrote: > How about opening a case with Redhat? > > It might help speed up the process.... > > /Yossi/ > > -----Original Message----- > From: rhelv6-list-bounces at redhat.com [mailto:rhelv6-list-bounces at redhat.com] On Behalf Of Brian Long (brilong) > Sent: Monday, February 13, 2012 4:31 PM > To: Red Hat Enterprise Linux 6 (Santiago) discussion mailing-list > Subject: Re: [rhelv6-list] WARNING: at net/sched/sch_generic.c:261 dev_watchdog leads to unresponsive network > > Could the system have old firmware on that NIC? Does your vendor provide firmware updates which could be installed? Without looking into the problem, I wonder if it's a driver/firmware interaction issue. > > /Brian/ > > On Feb 13, 2012, at 9:11 AM, daryl herzmann wrote: > >> Greetings all, >> >> I recently (~ 1 Jan) installed RHEL6 64bit on a system that has a: >> >> Intel Corporation 82574L Gigabit Network Connection >> >> which uses the e1000e module. Anyway, I've had nothing but trouble since. The network device becomes unresponsive after some seemingly random amount of time and a reboot is necessary to get it back functioning (reloading the module does not seem to help). I filed this issue in bugzilla here: >> >> https://bugzilla.redhat.com/show_bug.cgi?id=782586 >> >> The standard error that comes to the console is: >> >> Jan 17 12:38:46 xxx kernel: WARNING: at net/sched/sch_generic.c:261 >> dev_watchdog+0x26d/0x280() (Not tainted) Jan 17 12:38:46 xxx kernel: >> Hardware name: X9SCL/X9SCM Jan 17 12:38:46 xxx kernel: NETDEV >> WATCHDOG: eth0 (e1000e): transmit queu e 0 timed out >> >> This system is fully updated RHEL6.2. Anybody seem this before / have >> ideas on what to try to make it stop! :) >> >> thanks, >> daryl >> >> _______________________________________________ >> rhelv6-list mailing list >> rhelv6-list at redhat.com >> https://www.redhat.com/mailman/listinfo/rhelv6-list > > > _______________________________________________ > rhelv6-list mailing list > rhelv6-list at redhat.com > https://www.redhat.com/mailman/listinfo/rhelv6-list > > > > _______________________________________________ > rhelv6-list mailing list > rhelv6-list at redhat.com > https://www.redhat.com/mailman/listinfo/rhelv6-list -- +-----------------------------[ robin at redhat.com ]----+ | Robin Price II - RHCE,RHCDS,RHCVA | | Inside Solutions Architect | | Red Hat, Inc. | | w: +1 (919) 754 4412 | | c: +1 (252) 474 3525 | | | +---------[ http://people.redhat.com/rprice ]---------+ From akrherz at iastate.edu Mon Feb 13 16:26:40 2012 From: akrherz at iastate.edu (daryl herzmann) Date: Mon, 13 Feb 2012 10:26:40 -0600 (CST) Subject: [rhelv6-list] WARNING: at net/sched/sch_generic.c:261 dev_watchdog leads to unresponsive network In-Reply-To: <0LZC00JI39U3QBC0@davar.emet.co.il> References: <0LZC00JI39U3QBC0@davar.emet.co.il> Message-ID: On Mon, 13 Feb 2012, Yossi Goltz wrote: > How about opening a case with Redhat? > > It might help speed up the process.... Thanks, but we don't have that level of support with our academic site subscription. Our only options are bugzilla and to whine on email lists. :) daryl From amyagi at gmail.com Mon Feb 13 16:44:05 2012 From: amyagi at gmail.com (Akemi Yagi) Date: Mon, 13 Feb 2012 08:44:05 -0800 Subject: [rhelv6-list] WARNING: at net/sched/sch_generic.c:261 dev_watchdog leads to unresponsive network In-Reply-To: <4F39392A.7080000@redhat.com> References: <0LZC00JI39U3QBC0@davar.emet.co.il> <4F39392A.7080000@redhat.com> Message-ID: On Mon, Feb 13, 2012 at 8:24 AM, Robin Price II wrote: > I agree with Yossi here. ?Bugzilla is not a support tool for Red Hat and I > believe you would have better results opening a case. > > I did find a kbase on this matter for you: > > kernel: WARNING: at net/sched/sch_generic.c dev_watchdog() on RHEL6 > https://access.redhat.com/kb/docs/DOC-48625 > > Best of luck! > > ~rp As suggested by Brian Long earlier, and also in the referenced kbase, getting the latest firmware/driver may be worth a try. As for the driver, go to the ELRepo site: http://elrepo.org/tiki/Driver+Versions and see if there is a newer version than what is in RHEL-6. Akemi From bryan.hepworth at newcastle.ac.uk Wed Feb 15 09:56:17 2012 From: bryan.hepworth at newcastle.ac.uk (Bryan Hepworth) Date: Wed, 15 Feb 2012 09:56:17 +0000 Subject: [rhelv6-list] WARNING: at net/sched/sch_generic.c:261 dev_watchdog leads to unresponsive network In-Reply-To: References: Message-ID: <38ACA5EE9BB2474E8B3ECB1A38C5FB2CAF290AAA9E@EXSAN02.campus.ncl.ac.uk> > >Greetings all, > >I recently (~ 1 Jan) installed RHEL6 64bit on a system that has a: > > Intel Corporation 82574L Gigabit Network Connection > >which uses the e1000e module. Anyway, I've had nothing but trouble since. >The network device becomes unresponsive after some seemingly random >amount >of time and a reboot is necessary to get it back functioning (reloading >the module does not seem to help). I filed this issue in bugzilla here: > > https://bugzilla.redhat.com/show_bug.cgi?id=782586 > >The standard error that comes to the console is: > >Jan 17 12:38:46 xxx kernel: WARNING: at net/sched/sch_generic.c:261 >dev_watchdog+0x26d/0x280() (Not tainted) >Jan 17 12:38:46 xxx kernel: Hardware name: X9SCL/X9SCM >Jan 17 12:38:46 xxx kernel: NETDEV WATCHDOG: eth0 (e1000e): transmit queu >e 0 timed out > >This system is fully updated RHEL6.2. Anybody seem this before / have >ideas on what to try to make it stop! :) > Had something very similar to you, replaced the board and haven't had the problem since. Firmware update sounds like a plan. From rhelv6-list at redhat.com Tue Feb 21 14:54:08 2012 From: rhelv6-list at redhat.com (Red Hat Enterprise Linux 6 (Santiago) discussion mailing-list) Date: Tue, 21 Feb 2012 09:54:08 -0500 Subject: [rhelv6-list] Red Hat Enterprise Linux 5.8 Announcement Message-ID: <4F43B010.7020407@redhat.com> We?re pleased to announce the general availability of the Red Hat Enterprise Linux 5.8 operating system with new and improved functionality for enhanced performance, flexibility, and security. Our commitment to continuous innovation enables Red Hat Enterprise Linux to remain a scalable and trusted data center platform accommodating mission-critical enterprise workloads in physical, virtual, and cloud IT environments. And, the extension of the Red Hat Enterprise Linux 5 and 6 life cycle to 10 years means that our customers can plan IT deployments and migrations according to their business needs, not ours. Red Hat Enterprise Linux 5.8 supports new hardware platforms and includes the following enhancements: Hardware Support and Streamlined Installation --------------------------------------------- * Support for Power Management Quality of Service (QoS) provides automatic power management policy at the device level for managing latency and throughput based on QoS needs. During critical production windows, system administrators can enforce a high-performance, low-latency mode. * New iotop support enables the monitoring of I/O resources, such as storage devices, at the process level, providing insight into potential performance issues. * PCI-e 3.0 enablement provides support for new and emerging PCI-e 3.0 adapters. * Network, storage, and graphic driver enhancements are provided for new hardware support. * The configuration of IP over Infiniband (IPoIB) adapters is now supported by the system installer, making installation of Infiniband adapters easier. Virtualization -------------- * KVM scalability enhancements have increased the maximum supported virtual guest size from 128 to 256 virtual CPUs, extending the reach of virtualization deployments to large-scale workloads. * The spice-client package now adds support for Red Hat Enterprise Virtualization 3.0 and Red Hat Enterprise Linux 6.2 hosts. These improvements also allow users to access their desktop virtual machines over a WAN connection in addition to the local network. * KVM enhancements include improved guest boot times and improved clock and timer support. The KVM hypervisor has updated real time clock (RTC) support to improve the performance of Red Hat Enterprise Linux 6 guests on Red Hat Enterprise Virtualization or Red Hat Enterprise Linux 5 KVM hosts. * Additional KVM improvements increase the manageability and stability of KVM hosts and Red Hat Enterprise Linux guests. Many of these enhancements are in support of Red Hat Enterprise Virtualization 3.0. * Xen enhancements provide improved guest performance, improved logging for debug, and virtual disk re-sizing while a guest is running. Security Management ------------------- * OpenSCAP (Open Security Content Automation Protocol) support is now compliant with the SCAP 1.1 (Security Content Automation Protocol) specification, the most recent standard framework defined by National Institute of Standards and Technology (NIST) for creating a standardized approach for maintaining secure systems. * Red Hat Enterprise Linux 5 has been tested to the United States Government Configuration Board (USGCB) baseline and now meets the USGCB recommendations for government agencies following USGCB guidance for secure configuration policies. Subscription Asset Manager (SAM) Beta Availability -------------------------------------------------- We continue to invest in applications and functionality that provide our customers more flexibility and control over their Red Hat Enterprise Linux deployments. Accordingly, we are pleased to announce the Beta availability of the Red Hat Subscription Asset Manager (SAM). SAM is an asset tracking and inventory management application designed for on-premise deployment. SAM is the newest component in the Red Hat Entitlement offerings and will be included as part of your Red Hat Enterprise Linux subscription. For more information about Red Hat Enterprise Linux 5.8, the expanded Red Hat Enterprise Linux life cycle, and the Subscription Asset Manager, please visit: * Red Hat Enterprise Linux 5.x release notes: http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/5/ * Red Hat Enterprise Linux 10 year life cycle: http://www.redhat.com/rhel/lifecycle * SAM Get Started Guide and beta download: https://access.redhat.com/knowledge/solutions/70584 From gianluca.cecchi at gmail.com Wed Feb 22 17:56:42 2012 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Wed, 22 Feb 2012 18:56:42 +0100 Subject: [rhelv6-list] testing control groups... how to reset a previously set value limit? Message-ID: Hello, I'm testing some control groups features on a rh el 6.2 host as hypervisor and their effects inside one of its vms named testvm vm insists on some multipath devices on host, so 253:N devices. I want to limit I/O and run: echo "253:0 5000000" > /cgroup/blkio/libvirt/qemu/testvm/blkio.throttle.write_bps_device echo "253:1 5000000" > /cgroup/blkio/libvirt/qemu/testvm/blkio.throttle.write_bps_device echo "253:2 5000000" > /cgroup/blkio/libvirt/qemutestvm/blkio.throttle.write_bps_device echo "253:3 5000000" > /cgroup/blkio/libvirt/qemu/testvm/blkio.throttle.write_bps_device echo "253:5 5000000" > /cgroup/blkio/libvirt/qemu/testvm/blkio.throttle.write_bps_device echo "253:6 5000000" > /cgroup/blkio/libvirt/qemu/testvm/blkio.throttle.write_bps_device On hypervisor: # cgget -r blkio.throttle.read_bps_device libvirt/qemu/testvm libvirt/qemu/testvm: blkio.throttle.read_bps_device: 253:6 5000000 253:5 5000000 253:3 5000000 253:2 5000000 253:1 5000000 253:0 5000000 Now if I want to reset the values for blkio.throttle.write_bps_device to their original ones (so no limit), to have # cat /cgroup/blkio/libvirt/qemu/testvm/blkio.throttle.write_bps_device that returns no value, which syntax can I use? Thanks in advance, Gianluca From thomas.cameron at camerontech.com Thu Feb 23 00:04:06 2012 From: thomas.cameron at camerontech.com (Thomas Cameron) Date: Wed, 22 Feb 2012 18:04:06 -0600 Subject: [rhelv6-list] Opinions on multi-head Message-ID: <4F458276.30009@camerontech.com> Howdy All - I currently use a dual-head setup using a relatively inexpensive nVidia Corporation G73 [GeForce 7300 GT] card. I use the DVI and the VGA cables at the same time, and it works just fine. I can do wobbly windows and spinny desktops and all that jazz (yeah, I like the eye candy, sue me). I want to go to triple head. I am reasonably sure I can do it with e.g. an NVidia Quadro NVS 450, but I am curious as to whether the driver will do accelerated X? Three monitors will be nice, but not if I lose capabilities I have now with dual. So... anyone got any experience with triple-head or beyond? Any advice on NVidia Quadro vs. maybe ATI's offering? I'd love to hear any stories - good or bad - about accelerated X on more than two monitors. Thomas From jfranz at freerun.com Thu Feb 23 01:18:13 2012 From: jfranz at freerun.com (Jerry Franz) Date: Wed, 22 Feb 2012 17:18:13 -0800 Subject: [rhelv6-list] Opinions on multi-head In-Reply-To: <4F458276.30009@camerontech.com> References: <4F458276.30009@camerontech.com> Message-ID: <4F4593D5.7040406@freerun.com> On 02/22/2012 04:04 PM, Thomas Cameron wrote: > Howdy All - > > I currently use a dual-head setup using a relatively inexpensive > nVidia Corporation G73 [GeForce 7300 GT] card. I use the DVI and the > VGA cables at the same time, and it works just fine. I can do wobbly > windows and spinny desktops and all that jazz (yeah, I like the eye > candy, sue me). > > I want to go to triple head. I am reasonably sure I can do it with > e.g. an NVidia Quadro NVS 450, but I am curious as to whether the > driver will do accelerated X? Three monitors will be nice, but not if > I lose capabilities I have now with dual. > > So... anyone got any experience with triple-head or beyond? Any advice > on NVidia Quadro vs. maybe ATI's offering? > > I'd love to hear any stories - good or bad - about accelerated X on > more than two monitors. I use Ubuntu 11.10 for my desktop, but that said, I run an ATI 5770 card in triple head mode with full acceleration (and compiz, mainly to make Docky happy). All three run DVI with two coming from the built-in DVI ports and the third coming from an active DisplayPort -> DVI adapter. It took a bit of fiddling around with the Catalyst control panel to get working right, but the only really persistent problem I've had is that when you first log in the DisplayPort -> DVI doesn't always come up immediately. It can take a few minutes to decide to come on. Once it finally comes up it is fine. -- Benjamin Franz From jussi_rhel6 at silvennoinen.net Thu Feb 23 06:37:27 2012 From: jussi_rhel6 at silvennoinen.net (Jussi Silvennoinen) Date: Thu, 23 Feb 2012 08:37:27 +0200 (EET) Subject: [rhelv6-list] Opinions on multi-head In-Reply-To: <4F458276.30009@camerontech.com> References: <4F458276.30009@camerontech.com> Message-ID: > I currently use a dual-head setup using a relatively inexpensive nVidia > Corporation G73 [GeForce 7300 GT] card. I use the DVI and the VGA cables at > the same time, and it works just fine. I can do wobbly windows and spinny > desktops and all that jazz (yeah, I like the eye candy, sue me). > > I want to go to triple head. I am reasonably sure I can do it with e.g. an > NVidia Quadro NVS 450, but I am curious as to whether the driver will do > accelerated X? Three monitors will be nice, but not if I lose > capabilities I have now with dual. > > So... anyone got any experience with triple-head or beyond? Any advice on > NVidia Quadro vs. maybe ATI's offering? > > I'd love to hear any stories - good or bad - about accelerated X on more than > two monitors. NVS 290 worked fine with dualhead via DVI, however it's 3d performance was too sluggish for me so had to swap it for a Quadro 600. It gets around 9000 fps with glxgears compared to 2500 fps on the NVS 290. With the NVS 290, enabling compiz slowed down OpenGL stuff so much that I just couldn't live with it. User experience is silky smooth now. The 600 card I got had one DisplayPort and one DVI so I had to venture with a DP->DVI dongle. No problems there. Haven't found myself wanting triplehead yet. I've practically always had a nVidia in my workstation. No clue about the current state of AMD GPU support in Linux these days. -- Jussi From haiwu.us at gmail.com Thu Feb 23 09:30:22 2012 From: haiwu.us at gmail.com (hai wu) Date: Thu, 23 Feb 2012 03:30:22 -0600 Subject: [rhelv6-list] App install in RHEL6 KVM physical host Message-ID: This is not about RHEV, just plain KVM. Are there any concerns when installing heavy app in RHEL6 KVM physical host? -------------- next part -------------- An HTML attachment was scrubbed... URL: From goetz.reinicke at filmakademie.de Thu Feb 23 11:58:39 2012 From: goetz.reinicke at filmakademie.de (=?ISO-8859-15?Q?G=F6tz_Reinicke?=) Date: Thu, 23 Feb 2012 12:58:39 +0100 Subject: [rhelv6-list] invalid sender domain - postfix configuration question Message-ID: <4F4629EF.9070804@filmakademie.de> Hi, we do have a primary sendmail (rh el 5.x) server and some other new rh el 6.x servres which use that sendmail system as relay. I'm new to postfix and did not find the solution for the problem, that it seam that the mails send from the postfix systems are still nos some sort of masked by the mail.filmakademie.de sender address. For all mail originating from the postfix servers (relaied by teh sendmail server) the receiving servers complain about 'Deferred: 421 invalid sender domain' ... (misconfigured dns?) Can someone point me to the right config option? Thanks and best regards . G?tz -- G?tz Reinicke IT-Koordinator Tel. +49 7141 969 420 Fax +49 7141 969 55 420 E-Mail goetz.reinicke at filmakademie.de Filmakademie Baden-W?rttemberg GmbH Akademiehof 10 71638 Ludwigsburg www.filmakademie.de Eintragung Amtsgericht Stuttgart HRB 205016 Vorsitzender des Aufsichtsrats: J?rgen Walter MdL Staatssekret?r im Ministerium f?r Wissenschaft, Forschung und Kunst Baden-W?rttemberg Gesch?ftsf?hrer: Prof. Thomas Schadt -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 5161 bytes Desc: S/MIME Kryptografische Unterschrift URL: From gianluca.cecchi at gmail.com Thu Feb 23 12:13:36 2012 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Thu, 23 Feb 2012 13:13:36 +0100 Subject: [rhelv6-list] testing control groups... how to reset a previously set value limit? In-Reply-To: References: Message-ID: On Wed, Feb 22, 2012 at 6:56 PM, Gianluca Cecchi wrote: > On hypervisor: > # cgget -r blkio.throttle.read_bps_device libvirt/qemu/testvm > libvirt/qemu/testvm: > blkio.throttle.read_bps_device: 253:6 ? 5000000 > ? ? ? ?253:5 ? 5000000 > ? ? ? ?253:3 ? 5000000 > ? ? ? ?253:2 ? 5000000 > ? ? ? ?253:1 ? 5000000 > ? ? ? ?253:0 ? 5000000 > > Now if I want to reset the values for blkio.throttle.write_bps_device > to their original ones (so no limit), to have > > # cat /cgroup/blkio/libvirt/qemu/testvm/blkio.throttle.write_bps_device > > that returns no value, which syntax can I use? > > Thanks in advance, > Gianluca Ok, I found that for the blkio controller one has to put 0 as the value to reset SO to reset the limit for write bandwidth for dm-0 device: echo "253:0 0" > /cgroup/blkio/libvirt/qemu/testvm/blkio.throttle.write_bps_device or cgset -r blkio.throttle.write_bps_device="253:0 0" libvirt/qemu/testvm Gianluca From thomas.cameron at camerontech.com Thu Feb 23 13:22:28 2012 From: thomas.cameron at camerontech.com (Thomas Cameron) Date: Thu, 23 Feb 2012 07:22:28 -0600 Subject: [rhelv6-list] App install in RHEL6 KVM physical host In-Reply-To: References: Message-ID: <4F463D94.8000909@camerontech.com> On 02/23/2012 03:30 AM, hai wu wrote: > This is not about RHEV, just plain KVM. > > Are there any concerns when installing heavy app in RHEL6 KVM physical > host? There are considerations, of course. You can get as granular or as broad with tuning as you like. You can do things like use numactl and cpu pinning and control groups and sriov to really drive as much performance to the VM as possible. You want to do things like use virtio instead IDE for storage and network, things like that. Without more info about what you are going to do, it's really hard to give you a real recommendation. What is the workload you're thinking of virtualizing? I've got one customer who is running numerous Windows 2008R2 machines running workloads from file and print to Exchange to Sharepoint as KVM guests, and it's working just fine. TC From haiwu.us at gmail.com Thu Feb 23 14:03:04 2012 From: haiwu.us at gmail.com (hai wu) Date: Thu, 23 Feb 2012 08:03:04 -0600 Subject: [rhelv6-list] App install in RHEL6 KVM physical host In-Reply-To: <4F463D94.8000909@camerontech.com> References: <4F463D94.8000909@camerontech.com> Message-ID: My question was probably not clear, I was asking for app install on top of kvm's 'DOM0' physical (not a right word here, similar to xen dom0), such as cpu intensive puppet app, then in kvm guest, we could put lightweight things like apache, ldap servers, etc.. On 2/23/12, Thomas Cameron wrote: > On 02/23/2012 03:30 AM, hai wu wrote: >> This is not about RHEV, just plain KVM. >> >> Are there any concerns when installing heavy app in RHEL6 KVM physical >> host? > > There are considerations, of course. You can get as granular or as broad > with tuning as you like. You can do things like use numactl and cpu > pinning and control groups and sriov to really drive as much performance > to the VM as possible. You want to do things like use virtio instead IDE > for storage and network, things like that. > > Without more info about what you are going to do, it's really hard to > give you a real recommendation. What is the workload you're thinking of > virtualizing? > > I've got one customer who is running numerous Windows 2008R2 machines > running workloads from file and print to Exchange to Sharepoint as KVM > guests, and it's working just fine. > > TC > > _______________________________________________ > rhelv6-list mailing list > rhelv6-list at redhat.com > https://www.redhat.com/mailman/listinfo/rhelv6-list > -- Sent from my mobile device From thomas at redhat.com Thu Feb 23 14:20:24 2012 From: thomas at redhat.com (thomas at redhat.com) Date: Thu, 23 Feb 2012 08:20:24 -0600 Subject: [rhelv6-list] App install in RHEL6 KVM physical host In-Reply-To: References: <4F463D94.8000909@camerontech.com> Message-ID: <4F464B28.6030206@redhat.com> So, you want to install a high-impact application on the OS running on bare metal, and then run some lighter weight apps within virtual machines on the same hardware? That is easily done, but can be a bit cumbersome. It might be less complex to just run all your apps on one physical machine and use control groups to limit the lower impact apps than to bring up multiple OS instances. Thomas On 02/23/2012 08:03 AM, hai wu wrote: > My question was probably not clear, I was asking for app install on > top of kvm's 'DOM0' physical (not a right word here, similar to xen > dom0), such as cpu intensive puppet app, then in kvm guest, we could > put lightweight things like apache, ldap servers, etc.. > > > On 2/23/12, Thomas Cameron wrote: >> On 02/23/2012 03:30 AM, hai wu wrote: >>> This is not about RHEV, just plain KVM. >>> >>> Are there any concerns when installing heavy app in RHEL6 KVM physical >>> host? >> >> There are considerations, of course. You can get as granular or as broad >> with tuning as you like. You can do things like use numactl and cpu >> pinning and control groups and sriov to really drive as much performance >> to the VM as possible. You want to do things like use virtio instead IDE >> for storage and network, things like that. >> >> Without more info about what you are going to do, it's really hard to >> give you a real recommendation. What is the workload you're thinking of >> virtualizing? >> >> I've got one customer who is running numerous Windows 2008R2 machines >> running workloads from file and print to Exchange to Sharepoint as KVM >> guests, and it's working just fine. >> >> TC >> >> _______________________________________________ >> rhelv6-list mailing list >> rhelv6-list at redhat.com >> https://www.redhat.com/mailman/listinfo/rhelv6-list >> > -- Thomas Cameron, RHCA, RHCSS, RHCDS, RHCVA, RHCX Chief Architect, Canada and Central US 512-241-0774 office / 512-585-5631 cell http://people.redhat.com/tcameron/ IRC: choirboy / AIM: rhelguy / Yahoo: rhce_guy /Google+ http://ongpl.us/tdc From haiwu.us at gmail.com Thu Feb 23 14:50:06 2012 From: haiwu.us at gmail.com (hai wu) Date: Thu, 23 Feb 2012 08:50:06 -0600 Subject: [rhelv6-list] App install in RHEL6 KVM physical host In-Reply-To: <4F464B28.6030206@redhat.com> References: <4F463D94.8000909@camerontech.com> <4F464B28.6030206@redhat.com> Message-ID: Yes, but this way the guests app needs to run on top of the same OS release, which is not good.. On 2/23/12, thomas at redhat.com wrote: > So, you want to install a high-impact application on the OS running on > bare metal, and then run some lighter weight apps within virtual > machines on the same hardware? That is easily done, but can be a bit > cumbersome. > > It might be less complex to just run all your apps on one physical > machine and use control groups to limit the lower impact apps than to > bring up multiple OS instances. > > Thomas > > On 02/23/2012 08:03 AM, hai wu wrote: >> My question was probably not clear, I was asking for app install on >> top of kvm's 'DOM0' physical (not a right word here, similar to xen >> dom0), such as cpu intensive puppet app, then in kvm guest, we could >> put lightweight things like apache, ldap servers, etc.. >> >> >> On 2/23/12, Thomas Cameron wrote: >>> On 02/23/2012 03:30 AM, hai wu wrote: >>>> This is not about RHEV, just plain KVM. >>>> >>>> Are there any concerns when installing heavy app in RHEL6 KVM physical >>>> host? >>> >>> There are considerations, of course. You can get as granular or as broad >>> with tuning as you like. You can do things like use numactl and cpu >>> pinning and control groups and sriov to really drive as much performance >>> to the VM as possible. You want to do things like use virtio instead IDE >>> for storage and network, things like that. >>> >>> Without more info about what you are going to do, it's really hard to >>> give you a real recommendation. What is the workload you're thinking of >>> virtualizing? >>> >>> I've got one customer who is running numerous Windows 2008R2 machines >>> running workloads from file and print to Exchange to Sharepoint as KVM >>> guests, and it's working just fine. >>> >>> TC >>> >>> _______________________________________________ >>> rhelv6-list mailing list >>> rhelv6-list at redhat.com >>> https://www.redhat.com/mailman/listinfo/rhelv6-list >>> >> > > > -- > Thomas Cameron, RHCA, RHCSS, RHCDS, RHCVA, RHCX > Chief Architect, Canada and Central US > 512-241-0774 office / 512-585-5631 cell > http://people.redhat.com/tcameron/ > IRC: choirboy / AIM: rhelguy / Yahoo: rhce_guy /Google+ http://ongpl.us/tdc > > _______________________________________________ > rhelv6-list mailing list > rhelv6-list at redhat.com > https://www.redhat.com/mailman/listinfo/rhelv6-list > -- Sent from my mobile device From gianluca.cecchi at gmail.com Thu Feb 23 15:06:00 2012 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Thu, 23 Feb 2012 16:06:00 +0100 Subject: [rhelv6-list] App install in RHEL6 KVM physical host In-Reply-To: References: <4F463D94.8000909@camerontech.com> <4F464B28.6030206@redhat.com> Message-ID: On Thu, Feb 23, 2012 at 3:50 PM, hai wu wrote: > Yes, but this way the guests app needs to run on top of the same OS > release, which is not good.. Just to understand the contest, what is the limitation of not running "such as cpu intensive puppet app" in a dedicated guest too then? Eventually dedicating more vcpus to it? Lack of certification in virtual environments or what? Gianluca From prentice at ias.edu Thu Feb 23 16:50:03 2012 From: prentice at ias.edu (Prentice Bisbal) Date: Thu, 23 Feb 2012 11:50:03 -0500 Subject: [rhelv6-list] Opinions on multi-head In-Reply-To: <4F458276.30009@camerontech.com> References: <4F458276.30009@camerontech.com> Message-ID: <4F466E3B.5080700@ias.edu> On 02/22/2012 07:04 PM, Thomas Cameron wrote: > Howdy All - > > I currently use a dual-head setup using a relatively inexpensive > nVidia Corporation G73 [GeForce 7300 GT] card. I use the DVI and the > VGA cables at the same time, and it works just fine. I can do wobbly > windows and spinny desktops and all that jazz (yeah, I like the eye > candy, sue me). > > I want to go to triple head. I am reasonably sure I can do it with > e.g. an NVidia Quadro NVS 450, but I am curious as to whether the > driver will do accelerated X? Three monitors will be nice, but not if > I lose capabilities I have now with dual. > > So... anyone got any experience with triple-head or beyond? Any advice > on NVidia Quadro vs. maybe ATI's offering? > > I'd love to hear any stories - good or bad - about accelerated X on > more than two monitors. > I wouldn't use anything but NVidia with Linux. Open source zealots complain that the NVidia drivers aren't open source, but who cares? NVidia gives them away for free, and they've been extremely supportive of Linux since day one. I used to support high-end 3D scientific visualization Linux workstations, and I always used Nvidia hardware. The performance is awesome, the driver provides lots of tweaking, and it's incredibly well documented! You will never go wrong using NVidia on Linux. I think Intel provides good Linux support for their video cards, but I have no first hand knowledge. ATI's support of Linux leaves something to be desired. Some people say the ATI support is good and open source, but after days of googling, I still couldn't tell with absolute certainty which driver was the correct one for the ATI chip I was working with. I recommend avoiding ATI altogether. As far as whether or not your accelerated effects would work across all the monitors, the number of monitors your using shouldn't make a difference as long as the driver/video card you are using supports those 3D affects. If it works with one, it will work with 2, 3, 4, .... I would not try to mix-and-match video cards and drivers, however. This shouldn't be an issue with the Quadro NVS 450, since it's designed to support up to 4 displays from a single card. -- Prentice From pkrizak at qualcomm.com Thu Feb 23 17:05:06 2012 From: pkrizak at qualcomm.com (Krizak, Paul) Date: Thu, 23 Feb 2012 17:05:06 +0000 Subject: [rhelv6-list] Opinions on multi-head In-Reply-To: <4F466E3B.5080700@ias.edu> References: <4F458276.30009@camerontech.com> <4F466E3B.5080700@ias.edu> Message-ID: <8F8D3ACDBF996D458B659EA216D9E0E52113A0@nasanexd01f.na.qualcomm.com> I would hesitate to be quite so harsh on ATI graphics on Linux. I do agree that the drivers are less friendly than the nVidia drivers, but as long as you're using the binary blob drivers from AMD's website, they do work well. And the Catalyst Control Center has been ported to Linux and is included with the binary drivers, which makes setting up multi-head (even massive 6-display Eyefinity setups) trivially easy. ---- Paul Krizak Staff IT Engineer, IAS Qualcomm, Inc -----Original Message----- From: rhelv6-list-bounces at redhat.com [mailto:rhelv6-list-bounces at redhat.com] On Behalf Of Prentice Bisbal Sent: Thursday, February 23, 2012 8:50 AM To: rhelv6-list at redhat.com Subject: Re: [rhelv6-list] Opinions on multi-head On 02/22/2012 07:04 PM, Thomas Cameron wrote: > Howdy All - > > I currently use a dual-head setup using a relatively inexpensive > nVidia Corporation G73 [GeForce 7300 GT] card. I use the DVI and the > VGA cables at the same time, and it works just fine. I can do wobbly > windows and spinny desktops and all that jazz (yeah, I like the eye > candy, sue me). > > I want to go to triple head. I am reasonably sure I can do it with > e.g. an NVidia Quadro NVS 450, but I am curious as to whether the > driver will do accelerated X? Three monitors will be nice, but not if > I lose capabilities I have now with dual. > > So... anyone got any experience with triple-head or beyond? Any advice > on NVidia Quadro vs. maybe ATI's offering? > > I'd love to hear any stories - good or bad - about accelerated X on > more than two monitors. > I wouldn't use anything but NVidia with Linux. Open source zealots complain that the NVidia drivers aren't open source, but who cares? NVidia gives them away for free, and they've been extremely supportive of Linux since day one. I used to support high-end 3D scientific visualization Linux workstations, and I always used Nvidia hardware. The performance is awesome, the driver provides lots of tweaking, and it's incredibly well documented! You will never go wrong using NVidia on Linux. I think Intel provides good Linux support for their video cards, but I have no first hand knowledge. ATI's support of Linux leaves something to be desired. Some people say the ATI support is good and open source, but after days of googling, I still couldn't tell with absolute certainty which driver was the correct one for the ATI chip I was working with. I recommend avoiding ATI altogether. As far as whether or not your accelerated effects would work across all the monitors, the number of monitors your using shouldn't make a difference as long as the driver/video card you are using supports those 3D affects. If it works with one, it will work with 2, 3, 4, .... I would not try to mix-and-match video cards and drivers, however. This shouldn't be an issue with the Quadro NVS 450, since it's designed to support up to 4 displays from a single card. -- Prentice _______________________________________________ rhelv6-list mailing list rhelv6-list at redhat.com https://www.redhat.com/mailman/listinfo/rhelv6-list From haiwu.us at gmail.com Thu Feb 23 17:50:56 2012 From: haiwu.us at gmail.com (hai wu) Date: Thu, 23 Feb 2012 11:50:56 -0600 Subject: [rhelv6-list] App install in RHEL6 KVM physical host In-Reply-To: References: <4F463D94.8000909@camerontech.com> <4F464B28.6030206@redhat.com> Message-ID: There would be some overhead by putting apps into guest, by keeping app in the bare metal world, it should be faster .. I assume we could still do the same cgroup configuration in that case, dedicating more vcpus to the app as well? On 2/23/12, Gianluca Cecchi wrote: > On Thu, Feb 23, 2012 at 3:50 PM, hai wu wrote: >> Yes, but this way the guests app needs to run on top of the same OS >> release, which is not good.. > > Just to understand the contest, what is the limitation of not running > "such as cpu intensive puppet app" in a dedicated guest too then? > Eventually dedicating more vcpus to it? > Lack of certification in virtual environments or what? > > > Gianluca > > _______________________________________________ > rhelv6-list mailing list > rhelv6-list at redhat.com > https://www.redhat.com/mailman/listinfo/rhelv6-list > From pmeyer at themeyerfarm.com Thu Feb 23 20:39:54 2012 From: pmeyer at themeyerfarm.com (Phil Meyer) Date: Thu, 23 Feb 2012 13:39:54 -0700 Subject: [rhelv6-list] App install in RHEL6 KVM physical host In-Reply-To: References: <4F463D94.8000909@camerontech.com> <4F464B28.6030206@redhat.com> Message-ID: <4F46A41A.3090403@themeyerfarm.com> On 02/23/2012 10:50 AM, hai wu wrote: > There would be some overhead by putting apps into guest, by keeping > app in the bare metal world, it should be faster .. I assume we could > still do the same cgroup configuration in that case, dedicating more > vcpus to the app as well? > > On 2/23/12, Gianluca Cecchi wrote: >> On Thu, Feb 23, 2012 at 3:50 PM, hai wu wrote: >>> Yes, but this way the guests app needs to run on top of the same OS >>> release, which is not good.. >> Just to understand the contest, what is the limitation of not running >> "such as cpu intensive puppet app" in a dedicated guest too then? >> Eventually dedicating more vcpus to it? >> Lack of certification in virtual environments or what? >> The overhead of the hyper-visor piece is very minimal for almost everything. Graphics and real time i/o are still the big exceptions, but work is getting there to offer all types of PCI devices directly to the VM, but not just yet. Use dedicated drives or other storage space for VMs, pin them to CPUs, etc as has been noted. Here is your major issue, and one that cannot be ignored: What happens when you need to upgrade hardware (CPU/RAM). You are looking at some serious hours to move the OS to new hardware. Virtualize everything, and divorce all apps from the hardware. Need more? Bring up the new hardware and live migrate to it. No downtime at all. This justification alone eliminates almost all possible applications from needing bare Iron. Just use vitrio drivers, and you are always good to go (assuming intel->intel and AMD->AMD). You may also want to divorce all apps from specific storage requirements. This is harder, but should be in the plans for any large scale roll out of new servers. Think ISCSI, or Fibre Channel, or other reliable NAS before any consideration of local disks. A good storage array can way outperform a local disk. A 4GB RAM stick is cheaper than a hard drive dedicated to swap, so don't swap any more. And KVM/qemu is not going to force a need for swap on the server without you being able to calculate or see the need in advance. We are now heading back to the days of centralized services, but in many better ways. The concept is this: Migrate anything from anywhere to anywhere. For a good percentage of apps, we are there now. Good Luck! From ad+lists at uni-x.org Thu Feb 23 22:21:55 2012 From: ad+lists at uni-x.org (Alexander Dalloz) Date: Thu, 23 Feb 2012 23:21:55 +0100 Subject: [rhelv6-list] invalid sender domain - postfix configuration question In-Reply-To: <4F4629EF.9070804@filmakademie.de> References: <4F4629EF.9070804@filmakademie.de> Message-ID: <4F46BC03.1040907@uni-x.org> Am 23.02.2012 12:58, schrieb G?tz Reinicke: > Hi, > > we do have a primary sendmail (rh el 5.x) server and some other new rh > el 6.x servres which use that sendmail system as relay. > > I'm new to postfix and did not find the solution for the problem, that > it seam that the mails send from the postfix systems are still nos some > sort of masked by the mail.filmakademie.de sender address. > > For all mail originating from the postfix servers (relaied by teh > sendmail server) the receiving servers complain about 'Deferred: 421 > invalid sender domain' ... (misconfigured dns?) > > Can someone point me to the right config option? > > Thanks and best regards . G?tz Unfortunately you do not provide any configuration information nor precise log content. I suspect the Sendmail relay does not have the domain name used by the Postfix hosts as MASQUERADE_DOMAIN; see class {M} output: echo '$=M' | /usr/lib/sendmail -bt Alexander From leonfauster at googlemail.com Fri Feb 24 11:23:29 2012 From: leonfauster at googlemail.com (Leon Fauster) Date: Fri, 24 Feb 2012 12:23:29 +0100 Subject: [rhelv6-list] distcache disabled in apache Message-ID: <19751329-F2D1-4F2A-905E-135CC97A29BC@googlemail.com> Hi, i noticed that the distcache support is disabled in httpd. the httpd rpm changelog shows * Tue Dec 8 2009 Joe Orton - 2.2.14-2 - drop distcache support Any ideas why? Is there an alternative? Thx LF From goetz.reinicke at filmakademie.de Fri Feb 24 12:58:55 2012 From: goetz.reinicke at filmakademie.de (=?ISO-8859-1?Q?G=F6tz_Reinicke?=) Date: Fri, 24 Feb 2012 13:58:55 +0100 Subject: [rhelv6-list] invalid sender domain - postfix configuration question In-Reply-To: <4F46BC03.1040907@uni-x.org> References: <4F4629EF.9070804@filmakademie.de> <4F46BC03.1040907@uni-x.org> Message-ID: <4F47898F.40307@filmakademie.de> Am 23.02.12 23:21, schrieb Alexander Dalloz: > Am 23.02.2012 12:58, schrieb G?tz Reinicke: >> Hi, >> >> we do have a primary sendmail (rh el 5.x) server and some other new rh >> el 6.x servres which use that sendmail system as relay. >> >> I'm new to postfix and did not find the solution for the problem, that >> it seam that the mails send from the postfix systems are still nos some >> sort of masked by the mail.filmakademie.de sender address. >> >> For all mail originating from the postfix servers (relaied by teh >> sendmail server) the receiving servers complain about 'Deferred: 421 >> invalid sender domain' ... (misconfigured dns?) >> >> Can someone point me to the right config option? >> >> Thanks and best regards . G?tz > > Unfortunately you do not provide any configuration information nor > precise log content. > > I suspect the Sendmail relay does not have the domain name used by the > Postfix hosts as MASQUERADE_DOMAIN; see class {M} output: > > echo '$=M' | /usr/lib/sendmail -bt Hi Alexander, thanks for your comment. The output from that echo command is empty. And in my sendmail configuration there are no MASQUERADE* options set at all. My 'problem' is, if configured our sendmail mail environment some long time ago and everything worked for our setup. Adding the postfix new servers made/require some changes. So my mailserver relaying knowledge got a little bit rusty ... Basically I use the default configurations provided by redhat. So may be you could give me some mor advises? Thanks a log and best regards G?tz -- G?tz Reinicke IT-Koordinator Tel. +49 7141 969 420 Fax +49 7141 969 55 420 E-Mail goetz.reinicke at filmakademie.de Filmakademie Baden-W?rttemberg GmbH Akademiehof 10 71638 Ludwigsburg www.filmakademie.de Eintragung Amtsgericht Stuttgart HRB 205016 Vorsitzender des Aufsichtsrats: J?rgen Walter MdL Staatssekret?r im Ministerium f?r Wissenschaft, Forschung und Kunst Baden-W?rttemberg Gesch?ftsf?hrer: Prof. Thomas Schadt -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 5161 bytes Desc: S/MIME Kryptografische Unterschrift URL: From goetz.reinicke at filmakademie.de Fri Feb 24 13:46:33 2012 From: goetz.reinicke at filmakademie.de (=?ISO-8859-1?Q?G=F6tz_Reinicke?=) Date: Fri, 24 Feb 2012 14:46:33 +0100 Subject: [rhelv6-list] invalid sender domain - postfix configuration question - SOLVED In-Reply-To: <4F47898F.40307@filmakademie.de> References: <4F4629EF.9070804@filmakademie.de> <4F46BC03.1040907@uni-x.org> <4F47898F.40307@filmakademie.de> Message-ID: <4F4794B9.9060703@filmakademie.de> @-) the problematic system was a clone of a testsystem and the relaying for that hosts NEW IP was not added to the access file on the main sendmail server. Never run a changing system ... Sorry for the message noice on the list. Regards G?tz -- G?tz Reinicke IT-Koordinator Tel. +49 7141 969 420 Fax +49 7141 969 55 420 E-Mail goetz.reinicke at filmakademie.de Filmakademie Baden-W?rttemberg GmbH Akademiehof 10 71638 Ludwigsburg www.filmakademie.de Eintragung Amtsgericht Stuttgart HRB 205016 Vorsitzender des Aufsichtsrats: J?rgen Walter MdL Staatssekret?r im Ministerium f?r Wissenschaft, Forschung und Kunst Baden-W?rttemberg Gesch?ftsf?hrer: Prof. Thomas Schadt -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 5161 bytes Desc: S/MIME Kryptografische Unterschrift URL: From thomas.cameron at camerontech.com Fri Feb 24 15:08:40 2012 From: thomas.cameron at camerontech.com (Thomas Cameron) Date: Fri, 24 Feb 2012 09:08:40 -0600 Subject: [rhelv6-list] Opinions on multi-head In-Reply-To: <8F8D3ACDBF996D458B659EA216D9E0E52113A0@nasanexd01f.na.qualcomm.com> References: <4F458276.30009@camerontech.com> <4F466E3B.5080700@ias.edu> <8F8D3ACDBF996D458B659EA216D9E0E52113A0@nasanexd01f.na.qualcomm.com> Message-ID: <4F47A7F8.7050706@camerontech.com> Paul, do you have any specific card recommendations? I see Eyefinity listed on everything from cheap desktop class cards to massively expensive workstation cards. On 02/23/2012 11:05 AM, Krizak, Paul wrote: > I would hesitate to be quite so harsh on ATI graphics on Linux. I do agree that the drivers are less friendly than the nVidia drivers, but as long as you're using the binary blob drivers from AMD's website, they do work well. And the Catalyst Control Center has been ported to Linux and is included with the binary drivers, which makes setting up multi-head (even massive 6-display Eyefinity setups) trivially easy. > > ---- > Paul Krizak > Staff IT Engineer, IAS > Qualcomm, Inc > > -----Original Message----- > From: rhelv6-list-bounces at redhat.com [mailto:rhelv6-list-bounces at redhat.com] On Behalf Of Prentice Bisbal > Sent: Thursday, February 23, 2012 8:50 AM > To: rhelv6-list at redhat.com > Subject: Re: [rhelv6-list] Opinions on multi-head > > On 02/22/2012 07:04 PM, Thomas Cameron wrote: >> Howdy All - >> >> I currently use a dual-head setup using a relatively inexpensive >> nVidia Corporation G73 [GeForce 7300 GT] card. I use the DVI and the >> VGA cables at the same time, and it works just fine. I can do wobbly >> windows and spinny desktops and all that jazz (yeah, I like the eye >> candy, sue me). >> >> I want to go to triple head. I am reasonably sure I can do it with >> e.g. an NVidia Quadro NVS 450, but I am curious as to whether the >> driver will do accelerated X? Three monitors will be nice, but not if >> I lose capabilities I have now with dual. >> >> So... anyone got any experience with triple-head or beyond? Any advice >> on NVidia Quadro vs. maybe ATI's offering? >> >> I'd love to hear any stories - good or bad - about accelerated X on >> more than two monitors. >> > > I wouldn't use anything but NVidia with Linux. Open source zealots complain that the NVidia drivers aren't open source, but who cares? > NVidia gives them away for free, and they've been extremely supportive of Linux since day one. I used to support high-end 3D scientific visualization Linux workstations, and I always used Nvidia hardware. The performance is awesome, the driver provides lots of tweaking, and it's incredibly well documented! You will never go wrong using NVidia on Linux. > > I think Intel provides good Linux support for their video cards, but I have no first hand knowledge. > > ATI's support of Linux leaves something to be desired. Some people say the ATI support is good and open source, but after days of googling, I still couldn't tell with absolute certainty which driver was the correct one for the ATI chip I was working with. I recommend avoiding ATI altogether. > > As far as whether or not your accelerated effects would work across all the monitors, the number of monitors your using shouldn't make a difference as long as the driver/video card you are using supports those 3D affects. If it works with one, it will work with 2, 3, 4, .... I would not try to mix-and-match video cards and drivers, however. This shouldn't be an issue with the Quadro NVS 450, since it's designed to support up to 4 displays from a single card. > > -- > Prentice > > _______________________________________________ > rhelv6-list mailing list > rhelv6-list at redhat.com > https://www.redhat.com/mailman/listinfo/rhelv6-list > > _______________________________________________ > rhelv6-list mailing list > rhelv6-list at redhat.com > https://www.redhat.com/mailman/listinfo/rhelv6-list From michael.coffman at avagotech.com Fri Feb 24 16:07:38 2012 From: michael.coffman at avagotech.com (Michael Coffman) Date: Fri, 24 Feb 2012 09:07:38 -0700 Subject: [rhelv6-list] Changing login display manager Message-ID: Is there a way to get to a login screen that resembles what was available on rhel5? I would like to have login screen that allows the user to choose different window managers at the login. As a side note, I tried updating /etc/sysconfig/desktop. XDM works but gives no WM selection possibility. KDE fails miserably with kdmgreet crashing. the following is from /var/log/kdm.log: (EE) [drm] failed to open device (EE) open /dev/fb0: No such device error setting MTRR (base = 0xf5000000, size = 0x00e00000, type = 1) Invalid argument (22) expected keysym, got XF86TouchpadOn: line 120 of inet expected keysym, got XF86TouchpadOff: line 121 of inet expected keysym, got XF86TouchpadOn: line 120 of inet expected keysym, got XF86TouchpadOff: line 121 of inet KCrash: Application 'kdmgreet' crashing... error setting MTRR (base = 0xf5000000, size = 0x00e00000, type = 1) Invalid argument (22) Thanks for any pointers or info. -- -MichaelC From pkrizak at qualcomm.com Fri Feb 24 16:57:03 2012 From: pkrizak at qualcomm.com (Krizak, Paul) Date: Fri, 24 Feb 2012 16:57:03 +0000 Subject: [rhelv6-list] Opinions on multi-head In-Reply-To: <4F47A7F8.7050706@camerontech.com> References: <4F458276.30009@camerontech.com> <4F466E3B.5080700@ias.edu> <8F8D3ACDBF996D458B659EA216D9E0E52113A0@nasanexd01f.na.qualcomm.com> <4F47A7F8.7050706@camerontech.com> Message-ID: <8F8D3ACDBF996D458B659EA216D9E0E52119C3@nasanexd01f.na.qualcomm.com> The FirePro V3900 is a nice dual-head entry level card: http://www.amd.com/us/products/workstation/graphics/ati-firepro-3d/v3900/Pages/v3900.aspx Stepping up to the V4900 you get thee displays: http://www.amd.com/us/products/workstation/graphics/ati-firepro-3d/v4900/Pages/v4900.aspx Pretty much anything that says it supports Eyefinity will work nicely with the latest Catalyst drivers. Note that the FirePro and FireGL cards use a different, more heavily tested driver than the Radeon cards. I've personally never tried getting Eyefinity to work in Linux on a Radeon card, but that doesn't mean it won't work. ---- Paul Krizak cube: AR.250A8 Staff IT Engineer, IAS desk: 858-651-2467 Qualcomm, Inc cell: 512-791-0686 -----Original Message----- From: rhelv6-list-bounces at redhat.com [mailto:rhelv6-list-bounces at redhat.com] On Behalf Of Thomas Cameron Sent: Friday, February 24, 2012 7:09 AM To: rhelv6-list at redhat.com Subject: Re: [rhelv6-list] Opinions on multi-head Paul, do you have any specific card recommendations? I see Eyefinity listed on everything from cheap desktop class cards to massively expensive workstation cards. On 02/23/2012 11:05 AM, Krizak, Paul wrote: > I would hesitate to be quite so harsh on ATI graphics on Linux. I do agree that the drivers are less friendly than the nVidia drivers, but as long as you're using the binary blob drivers from AMD's website, they do work well. And the Catalyst Control Center has been ported to Linux and is included with the binary drivers, which makes setting up multi-head (even massive 6-display Eyefinity setups) trivially easy. > > ---- > Paul Krizak > Staff IT Engineer, IAS > Qualcomm, Inc > > -----Original Message----- > From: rhelv6-list-bounces at redhat.com > [mailto:rhelv6-list-bounces at redhat.com] On Behalf Of Prentice Bisbal > Sent: Thursday, February 23, 2012 8:50 AM > To: rhelv6-list at redhat.com > Subject: Re: [rhelv6-list] Opinions on multi-head > > On 02/22/2012 07:04 PM, Thomas Cameron wrote: >> Howdy All - >> >> I currently use a dual-head setup using a relatively inexpensive >> nVidia Corporation G73 [GeForce 7300 GT] card. I use the DVI and the >> VGA cables at the same time, and it works just fine. I can do wobbly >> windows and spinny desktops and all that jazz (yeah, I like the eye >> candy, sue me). >> >> I want to go to triple head. I am reasonably sure I can do it with >> e.g. an NVidia Quadro NVS 450, but I am curious as to whether the >> driver will do accelerated X? Three monitors will be nice, but not >> if I lose capabilities I have now with dual. >> >> So... anyone got any experience with triple-head or beyond? Any >> advice on NVidia Quadro vs. maybe ATI's offering? >> >> I'd love to hear any stories - good or bad - about accelerated X on >> more than two monitors. >> > > I wouldn't use anything but NVidia with Linux. Open source zealots complain that the NVidia drivers aren't open source, but who cares? > NVidia gives them away for free, and they've been extremely supportive of Linux since day one. I used to support high-end 3D scientific visualization Linux workstations, and I always used Nvidia hardware. The performance is awesome, the driver provides lots of tweaking, and it's incredibly well documented! You will never go wrong using NVidia on Linux. > > I think Intel provides good Linux support for their video cards, but I have no first hand knowledge. > > ATI's support of Linux leaves something to be desired. Some people say the ATI support is good and open source, but after days of googling, I still couldn't tell with absolute certainty which driver was the correct one for the ATI chip I was working with. I recommend avoiding ATI altogether. > > As far as whether or not your accelerated effects would work across all the monitors, the number of monitors your using shouldn't make a difference as long as the driver/video card you are using supports those 3D affects. If it works with one, it will work with 2, 3, 4, .... I would not try to mix-and-match video cards and drivers, however. This shouldn't be an issue with the Quadro NVS 450, since it's designed to support up to 4 displays from a single card. > > -- > Prentice > > _______________________________________________ > rhelv6-list mailing list > rhelv6-list at redhat.com > https://www.redhat.com/mailman/listinfo/rhelv6-list > > _______________________________________________ > rhelv6-list mailing list > rhelv6-list at redhat.com > https://www.redhat.com/mailman/listinfo/rhelv6-list _______________________________________________ rhelv6-list mailing list rhelv6-list at redhat.com https://www.redhat.com/mailman/listinfo/rhelv6-list