Re: [ANN] oVirt 4.3.7 First Release Candidate is now available for testing
Hi Dominik, I think I found it - it seems that for some reason my gateway is missing in 'ip route'. Running 'ifdown ovirtmgmt; ifup ovirtmgmt' fixes it, which proves that network scripts read properly the ifcfg-ovirtmgmt file. Do you have any idea where to look for? Best Regards, Strahil NikolovOn Nov 2, 2019 17:17, Strahil <hunter86_bg@yahoo.com> wrote:
I think I have a theory about ovirt1 , but I have to test it.
1. Ovirt1 has a team device controlled by NetworkManager , but in engine those interfaces are unused 2. I guess that when I recently 'reinstalled' the host from UI, the old teamports' ifcfg-<DEVICENAME> was moved to ifcfg-<DEVICENAME>-1 and the original file was left empty. 3. When the system rebooted, network.service rereads all ifcfg-* files and starts the team0 , but when NetworkManager reads the empty files - he alters the /etc/resolv.conf which leads to out-of-sync status.
Still, this is only theory and as ovirt3 has no team device (plain gluster arbiter) - I'm not sure how close is this one.
Best Regards, Strahil Nikolov
On Nov 2, 2019 09:50, Strahil Nikolov <hunter86_bg@yahoo.com> wrote:
I saw the hint in the Cluster's tab and then on each host. I had to put the host back to maintenance in order to sync it again.
Logs attached .
Best Regards, Strahil Nikolov
В четвъртък, 31 октомври 2019 г., 16:12:53 ч. Гринуич+2, Dominik Holler <dholler@redhat.com> написа:
On Thu, Oct 31, 2019 at 10:06 AM Sandro Bonazzola <sbonazzo@redhat.com> wrote:
Il giorno gio 31 ott 2019 alle ore 09:45 Strahil Nikolov <hunter86_bg@yahoo.com> ha scritto:
Hi Sandro,
I have just updated to latest release candidate (2 clusters in total - 1 AMD , 1 Intel). The only issue observed is that after patching every host's ovirtmgmt (I have another one that has none of the 4 roles which was OK) is out of sync and has to be synced before removing maintenance.
Still the system was accessible.
Can you guide me which log reports the out-of-sync in order to further investigate.
+Miguel Duarte de Mora Barroso , +Dominik Holler can you please help here?
Do you remember if there was a hint shown as a tooltip if you hover the mouse pointer over ovirtmgmt in "Compute > Hosts > xxx > Network Interfaces > Setup Host Networks" ?
Are you able to share the vdsm.log and engine.log after the ovirtmgmt?
On Sat, Nov 2, 2019 at 4:59 PM Strahil <hunter86_bg@yahoo.com> wrote:
Hi Dominik,
I think I found it - it seems that for some reason my gateway is missing in 'ip route'. Running 'ifdown ovirtmgmt; ifup ovirtmgmt' fixes it, which proves that network scripts read properly the ifcfg-ovirtmgmt file.
Do you have any idea where to look for?
Does this refers to the host ovirt3 without the team interface? Maybe there is any other interface than ovirtmgmt which influences the default gw? Is there an indication in any other ifcfg file or is a connection managed by networkmanager?
Best Regards, Strahil Nikolov On Nov 2, 2019 17:17, Strahil <hunter86_bg@yahoo.com> wrote:
I think I have a theory about ovirt1 , but I have to test it.
1. Ovirt1 has a team device controlled by NetworkManager , but in engine those interfaces are unused
What is your motivation to use a team device, instead a bond managed by oVirt?
2. I guess that when I recently 'reinstalled' the host from UI, the old teamports' ifcfg-<DEVICENAME> was moved to ifcfg-<DEVICENAME>-1 and the original file was left empty. 3. When the system rebooted, network.service rereads all ifcfg-* files and starts the team0 , but when NetworkManager reads the empty files - he alters the /etc/resolv.conf which leads to out-of-sync status.
Still, this is only theory and as ovirt3 has no team device (plain gluster arbiter) - I'm not sure how close is this one.
Best Regards, Strahil Nikolov On Nov 2, 2019 09:50, Strahil Nikolov <hunter86_bg@yahoo.com> wrote:
I saw the hint in the Cluster's tab and then on each host. I had to put the host back to maintenance in order to sync it again.
Logs attached .
Best Regards, Strahil Nikolov
В четвъртък, 31 октомври 2019 г., 16:12:53 ч. Гринуич+2, Dominik Holler < dholler@redhat.com> написа:
On Thu, Oct 31, 2019 at 10:06 AM Sandro Bonazzola <sbonazzo@redhat.com> wrote:
Il giorno gio 31 ott 2019 alle ore 09:45 Strahil Nikolov < hunter86_bg@yahoo.com> ha scritto:
Hi Sandro,
I have just updated to latest release candidate (2 clusters in total - 1 AMD , 1 Intel). The only issue observed is that after patching every host's ovirtmgmt (I have another one that has none of the 4 roles which was OK) is out of sync and has to be synced before removing maintenance.
Still the system was accessible.
Can you guide me which log reports the out-of-sync in order to further investigate.
+Miguel Duarte de Mora Barroso <mduarted@redhat.com> , +Dominik Holler <dholler@redhat.com> can you please help here?
Do you remember if there was a hint shown as a tooltip if you hover the mouse pointer over ovirtmgmt in "Compute > Hosts > xxx > Network Interfaces > Setup Host Networks" ?
Are you able to share the vdsm.log and engine.log after the ovirtmgmt?
It's releated to all 3 nodes.I've created a vdsm hook in 'after_ifcfg_write' that sets a 'DELAY=5' and 'STP=on' and after reboot everything is OK. I have just completely powered off the lab and after powering on - the gateway is OK, so it cannot be reproduced any longer. Best Regards,Strahil Nikolov В понеделник, 4 ноември 2019 г., 10:12:47 ч. Гринуич+2, Dominik Holler <dholler@redhat.com> написа: On Sat, Nov 2, 2019 at 4:59 PM Strahil <hunter86_bg@yahoo.com> wrote: Hi Dominik, I think I found it - it seems that for some reason my gateway is missing in 'ip route'. Running 'ifdown ovirtmgmt; ifup ovirtmgmt' fixes it, which proves that network scripts read properly the ifcfg-ovirtmgmt file. Do you have any idea where to look for? Does this refers to the host ovirt3 without the team interface?Maybe there is any other interface than ovirtmgmt which influences the default gw? Is there an indication in any other ifcfg file or is a connection managed by networkmanager? Best Regards, Strahil Nikolov On Nov 2, 2019 17:17, Strahil <hunter86_bg@yahoo.com> wrote: I think I have a theory about ovirt1 , but I have to test it. 1. Ovirt1 has a team device controlled by NetworkManager , but in engine those interfaces are unused What is your motivation to use a team device, instead a bond managed by oVirt? 2. I guess that when I recently 'reinstalled' the host from UI, the old teamports' ifcfg-<DEVICENAME> was moved to ifcfg-<DEVICENAME>-1 and the original file was left empty. 3. When the system rebooted, network.service rereads all ifcfg-* files and starts the team0 , but when NetworkManager reads the empty files - he alters the /etc/resolv.conf which leads to out-of-sync status. Still, this is only theory and as ovirt3 has no team device (plain gluster arbiter) - I'm not sure how close is this one. Best Regards, Strahil Nikolov On Nov 2, 2019 09:50, Strahil Nikolov <hunter86_bg@yahoo.com> wrote: I saw the hint in the Cluster's tab and then on each host. I had to put the host back to maintenance in order to sync it again. Logs attached . Best Regards,Strahil Nikolov В четвъртък, 31 октомври 2019 г., 16:12:53 ч. Гринуич+2, Dominik Holler <dholler@redhat.com> написа: On Thu, Oct 31, 2019 at 10:06 AM Sandro Bonazzola <sbonazzo@redhat.com> wrote: Il giorno gio 31 ott 2019 alle ore 09:45 Strahil Nikolov <hunter86_bg@yahoo.com> ha scritto: Hi Sandro, I have just updated to latest release candidate (2 clusters in total - 1 AMD , 1 Intel). The only issue observed is that after patching every host's ovirtmgmt (I have another one that has none of the 4 roles which was OK) is out of sync and has to be synced before removing maintenance. Still the system was accessible. Can you guide me which log reports the out-of-sync in order to further investigate. +Miguel Duarte de Mora Barroso , +Dominik Holler can you please help here? Do you remember if there was a hint shown as a tooltip if you hover the mouse pointer over ovirtmgmt in "Compute > Hosts > xxx > Network Interfaces > Setup Host Networks" ? Are you able to share the vdsm.log and engine.log after the ovirtmgmt? _______________________________________________ Users mailing list -- users@ovirt.org To unsubscribe send an email to users-leave@ovirt.org Privacy Statement: https://www.ovirt.org/site/privacy-policy/ oVirt Code of Conduct: https://www.ovirt.org/community/about/community-guidelines/ List Archives: https://lists.ovirt.org/archives/list/users@ovirt.org/message/BL6H2R5FWY7CF2...
Best Regards, Strahil Nikolov On Nov 2, 2019 17:17, Strahil <hunter86_bg@yahoo.com> wrote: I think I have a theory about ovirt1 , but I have to test it. 1. Ovirt1 has a team device controlled by NetworkManager , but in engine those interfaces are unused
What is your motivation to use a team device, instead a bond managed by oVirt? This was already discussed in 'Is teamd supported on 4.3' in which you were already involved. I can't say it better -so I will quote:
Hi Dominik and thanks for the quick reply. Purely the load balancing for LACP support and the low overhead. Since teamd is probably going to be the future I wanted to have oVirt using it. Furthermore, teamd is much more flexible to extend, add new interfaces, change priorities and you name it.
It would be great if teamd support was added in the upcoming oVirt releases. I've seen there is an RFE for the same question I asked here.
Thanks Valentin for your nice explanation. Would you comment this on the RFE Bug 1351510 - [RFE] Support using Team devices instead of bond devices or may I copy your text there? Best Regards,Strahil Nikolov
participants (3)
-
Dominik Holler -
Strahil -
Strahil Nikolov