On 03/08/2012 02:48 AM, Sorin Srbu wrote:
Another smooth update. Thanks CentOS-team!
Thanks! Lots of very hard work by the QA team.
Everyone needs to look at their /var/log/messages if they have an SAS hard drive.
This is a very big (and potentially dangerous) problem:
On 03/08/12 3:33 AM, Johnny Hughes wrote:
On 03/08/2012 02:48 AM, Sorin Srbu wrote:
Another smooth update. Thanks CentOS-team!
Thanks! Lots of very hard work by the QA team.
Everyone needs to look at their /var/log/messages if they have an SAS hard drive.
This is a very big (and potentially dangerous) problem:
is this problem specific to the 1068E chipset, or is it likely that the MPT-SAS2 2008 cards (9210-8i, etc) have the same problem?
On 03/08/2012 06:37 AM, John R Pierce wrote:
On 03/08/12 3:33 AM, Johnny Hughes wrote:
On 03/08/2012 02:48 AM, Sorin Srbu wrote:
Another smooth update. Thanks CentOS-team!
Thanks! Lots of very hard work by the QA team.
Everyone needs to look at their /var/log/messages if they have an SAS hard drive.
This is a very big (and potentially dangerous) problem:
is this problem specific to the 1068E chipset, or is it likely that the MPT-SAS2 2008 cards (9210-8i, etc) have the same problem?
I have only personally seen it on that chipset ... but the Fedora bug seems to have it on other cards that use the mptsas driver.
Am 08.03.2012 12:33, schrieb Johnny Hughes:
On 03/08/2012 02:48 AM, Sorin Srbu wrote:
Another smooth update. Thanks CentOS-team!
Thanks! Lots of very hard work by the QA team.
Everyone needs to look at their /var/log/messages if they have an SAS hard drive.
This is a very big (and potentially dangerous) problem:
I can only second the warning about 5.8.
Apart from running kvm and virtio drivers in guest ceasing disk io: https://www.redhat.com/archives/rhelv5-list/2012-March/msg00017.html
we are seeing problems with NFS server crashing on 5.8.
Redhat has released a new kernel 2.6.18-308.1.1 yesterday, but I have not had the time to test this yet.
Rainer
On 8.3.2012 13:44, Rainer Traut wrote:
Am 08.03.2012 12:33, schrieb Johnny Hughes:
On 03/08/2012 02:48 AM, Sorin Srbu wrote:
Another smooth update. Thanks CentOS-team!
Thanks! Lots of very hard work by the QA team.
Everyone needs to look at their /var/log/messages if they have an SAS hard drive.
This is a very big (and potentially dangerous) problem:
I can only second the warning about 5.8.
Apart from running kvm and virtio drivers in guest ceasing disk io: https://www.redhat.com/archives/rhelv5-list/2012-March/msg00017.html
I read your original message regarding this https://www.redhat.com/archives/rhelv5-list/2012-February/msg00060.html
according to your experiences only upgrading kvm hosts are problematic? so upgrading only guests to 5.8 is maybe fine?
Am 08.03.2012 15:37, schrieb Markus Falb:
I read your original message regarding this https://www.redhat.com/archives/rhelv5-list/2012-February/msg00060.html
according to your experiences only upgrading kvm hosts are problematic? so upgrading only guests to 5.8 is maybe fine?
I have had limited time of testing so I'd take this with a grain of salt. After experiencing the NFS problem I rolled all the KVM hosts and guests back to 5.7 kernel and kvm module.
Rainer
On 9.3.2012 09:43, Rainer Traut wrote:
Am 08.03.2012 15:37, schrieb Markus Falb:
I read your original message regarding this https://www.redhat.com/archives/rhelv5-list/2012-February/msg00060.html
according to your experiences only upgrading kvm hosts are problematic? so upgrading only guests to 5.8 is maybe fine?
I have had limited time of testing so I'd take this with a grain of salt. After experiencing the NFS problem I rolled all the KVM hosts and guests back to 5.7 kernel and kvm module.
Actually it hit me too. 2 guests on the same hosts went unresponsive. The kernel on the host was not updated but kvm and kmod-kvm was. I try with downgraded kvm and kmod-kvm on the host and downgraded kernel on the guests like you did.
sigh
On 03/10/2012 08:12 PM, Markus Falb wrote:
On 9.3.2012 09:43, Rainer Traut wrote:
Am 08.03.2012 15:37, schrieb Markus Falb:
I read your original message regarding this https://www.redhat.com/archives/rhelv5-list/2012-February/msg00060.html
according to your experiences only upgrading kvm hosts are problematic? so upgrading only guests to 5.8 is maybe fine?
I have had limited time of testing so I'd take this with a grain of salt. After experiencing the NFS problem I rolled all the KVM hosts and guests back to 5.7 kernel and kvm module.
Actually it hit me too. 2 guests on the same hosts went unresponsive. The kernel on the host was not updated but kvm and kmod-kvm was. I try with downgraded kvm and kmod-kvm on the host and downgraded kernel on the guests like you did.
sigh
I upgraded the kernel to kernel-2.6.18-308.1.1.el5 (also latest kvm and libvirt) on my host on Thursday night and have had no issues for more than 2 days.
On 03/10/2012 10:57 PM, Johnny Hughes wrote:
On 03/10/2012 08:12 PM, Markus Falb wrote:
On 9.3.2012 09:43, Rainer Traut wrote:
Am 08.03.2012 15:37, schrieb Markus Falb:
I read your original message regarding this https://www.redhat.com/archives/rhelv5-list/2012-February/msg00060.html
according to your experiences only upgrading kvm hosts are problematic? so upgrading only guests to 5.8 is maybe fine?
I have had limited time of testing so I'd take this with a grain of salt. After experiencing the NFS problem I rolled all the KVM hosts and guests back to 5.7 kernel and kvm module.
Actually it hit me too. 2 guests on the same hosts went unresponsive. The kernel on the host was not updated but kvm and kmod-kvm was. I try with downgraded kvm and kmod-kvm on the host and downgraded kernel on the guests like you did.
sigh
I upgraded the kernel to kernel-2.6.18-308.1.1.el5 (also latest kvm and libvirt) on my host on Thursday night and have had no issues for more than 2 days.
And, now it has crashed. So whatever the issue is, the kernel-2.6.18-308.1.1.el5 does not fix it. As suggested on the RHEL5 list, I have shifted the problem VMs off of the virtio disk driver and to the standard driver. you do this as detailed in this mail:
https://www.redhat.com/archives/rhelv5-list/2012-March/msg00017.html
We will see if this fixes the issue.
I have created an upstream bugzilla entry here:
On 03/08/2012 06:44 AM, Rainer Traut wrote:
Am 08.03.2012 12:33, schrieb Johnny Hughes:
On 03/08/2012 02:48 AM, Sorin Srbu wrote:
Another smooth update. Thanks CentOS-team!
Thanks! Lots of very hard work by the QA team.
Everyone needs to look at their /var/log/messages if they have an SAS hard drive.
This is a very big (and potentially dangerous) problem:
I can only second the warning about 5.8.
Apart from running kvm and virtio drivers in guest ceasing disk io: https://www.redhat.com/archives/rhelv5-list/2012-March/msg00017.html
I have seen that KVM issue with virtio drives happen two times with the same CentOS-5.8 guest on a 5.8 host ... they happened about 48 hours apart. I have upgraded the host machine (and the guest) to the latest 5.8 kernel (kernel-2.6.18-308.1.1.el5.x86_64). I will post here if the issue recurs.
we are seeing problems with NFS server crashing on 5.8.
Redhat has released a new kernel 2.6.18-308.1.1 yesterday, but I have not had the time to test this yet.
-
John R Pierce -
Johnny Hughes -
Markus Falb -
Rainer Traut