Update of plain CentOS hosts very slow
Hello, I have a 4.4.5 environment that I'm upgrading to 4.4.6. I'm upgrading plain CentOS hosts from the GUI. They are in 4.4.5, so in particular CentOS 8.3 and as part of the upgrade they have to be put to 8.4. In the past I used "yum update" on the host but now it seems it is not the correct way. But the ansible part related to package updates seems to be very slow. It gives the impression that it is doing it one by one and not as a whole when you run "yum update" Now it is about 30 minutes that the update is going on and my internet speed is for sure very high. In messages of host I see every single line suche this ones: Jun 8 11:09:30 ov300 python3[3031815]: ansible-dnf Invoked with name=['rsyslog-relp.x86_64'] state=latest lock_timeout=300 conf_file=/tmp/yum.conf allow_downgrade=False autoremove=False bugfix=False disable_gpg_check=False disable_plugin=[] disablerepo=[] download_only=False enable_plugin=[] enablerepo=[] exclude=[] installroot=/ install_repoquery=True install_weak_deps=True security=False skip_broken=False update_cache=False update_only=False validate_certs=True disable_excludes=None download_dir=None list=None releasever=None Jun 8 11:09:32 ov300 python3[3031828]: ansible-dnf Invoked with name=['runc.x86_64'] state=latest lock_timeout=300 conf_file=/tmp/yum.conf allow_downgrade=False autoremove=False bugfix=False disable_gpg_check=False disable_plugin=[] disablerepo=[] download_only=False enable_plugin=[] enablerepo=[] exclude=[] installroot=/ install_repoquery=True install_weak_deps=True security=False skip_broken=False update_cache=False update_only=False validate_certs=True disable_excludes=None download_dir=None list=None releasever=None Any clarification? Thanks, Gianluca
On Tue, Jun 8, 2021 at 11:15 AM Gianluca Cecchi <gianluca.cecchi@gmail.com> wrote:
Hello, I have a 4.4.5 environment that I'm upgrading to 4.4.6.
I'm upgrading plain CentOS hosts from the GUI. They are in 4.4.5, so in particular CentOS 8.3 and as part of the upgrade they have to be put to 8.4.
In the past I used "yum update" on the host but now it seems it is not the correct way.
But the ansible part related to package updates seems to be very slow. It gives the impression that it is doing it one by one and not as a whole when you run "yum update" Now it is about 30 minutes that the update is going on and my internet speed is for sure very high.
In messages of host I see every single line suche this ones:
Jun 8 11:09:30 ov300 python3[3031815]: ansible-dnf Invoked with name=['rsyslog-relp.x86_64'] state=latest lock_timeout=300 conf_file=/tmp/yum.conf allow_downgrade=False autoremove=False bugfix=False disable_gpg_check=False disable_plugin=[] disablerepo=[] download_only=False enable_plugin=[] enablerepo=[] exclude=[] installroot=/ install_repoquery=True install_weak_deps=True security=False skip_broken=False update_cache=False update_only=False validate_certs=True disable_excludes=None download_dir=None list=None releasever=None Jun 8 11:09:32 ov300 python3[3031828]: ansible-dnf Invoked with name=['runc.x86_64'] state=latest lock_timeout=300 conf_file=/tmp/yum.conf allow_downgrade=False autoremove=False bugfix=False disable_gpg_check=False disable_plugin=[] disablerepo=[] download_only=False enable_plugin=[] enablerepo=[] exclude=[] installroot=/ install_repoquery=True install_weak_deps=True security=False skip_broken=False update_cache=False update_only=False validate_certs=True disable_excludes=None download_dir=None list=None releasever=None
Any clarification?
Thanks, Gianluca
BTW the update had a duration of 33 minutes (I selected to not reboot the host): Jun 8, 2021, 10:40:35 AM Host ov300 upgrade was started (User: tekka@mydomain). Jun 8, 2021, 11:13:13 AM Host ov300 upgrade was completed successfully. At the end if I open a terminal I see: [root@ov300 ~]# rpm -q ansible ansible-2.9.16-2.el8.noarch [root@ov300 ~]# and [root@ov300 ~]# yum update Last metadata expiration check: 0:39:50 ago on Tue 08 Jun 2021 10:41:09 AM CEST. Dependencies resolved. =============================================================================================================== Package Architecture Version Repository Size =============================================================================================================== Upgrading: ansible noarch 2.9.21-2.el8 ovirt-4.4-centos-ovirt44 17 M Transaction Summary =============================================================================================================== Upgrade 1 Package Total download size: 17 M Is this ok [y/N]: Operation aborted. [root@ov300 ~]# Is it expected? Currently in my yum.conf I have: [main] gpgcheck=1 installonly_limit=3 clean_requirements_on_remove=True best=True skip_if_unavailable=False Gianluca
Hi Gianluca, Upgrade host time depends on the upgrading that needs to be done As for your second question, we updated ansible only if there was another package which required a new version of it, and we have used a different (temporarily created) yum.conf file which had the setting best=False, so yes, this behavior is currently expected, (and will be changed as we're going to allow updating ansible package for el8 hosts too) Hope this helps, Dana On Tue, Jun 8, 2021 at 12:28 PM Gianluca Cecchi <gianluca.cecchi@gmail.com> wrote:
On Tue, Jun 8, 2021 at 11:15 AM Gianluca Cecchi <gianluca.cecchi@gmail.com> wrote:
Hello, I have a 4.4.5 environment that I'm upgrading to 4.4.6.
I'm upgrading plain CentOS hosts from the GUI. They are in 4.4.5, so in particular CentOS 8.3 and as part of the upgrade they have to be put to 8.4.
In the past I used "yum update" on the host but now it seems it is not the correct way.
But the ansible part related to package updates seems to be very slow. It gives the impression that it is doing it one by one and not as a whole when you run "yum update" Now it is about 30 minutes that the update is going on and my internet speed is for sure very high.
In messages of host I see every single line suche this ones:
Jun 8 11:09:30 ov300 python3[3031815]: ansible-dnf Invoked with name=['rsyslog-relp.x86_64'] state=latest lock_timeout=300 conf_file=/tmp/yum.conf allow_downgrade=False autoremove=False bugfix=False disable_gpg_check=False disable_plugin=[] disablerepo=[] download_only=False enable_plugin=[] enablerepo=[] exclude=[] installroot=/ install_repoquery=True install_weak_deps=True security=False skip_broken=False update_cache=False update_only=False validate_certs=True disable_excludes=None download_dir=None list=None releasever=None Jun 8 11:09:32 ov300 python3[3031828]: ansible-dnf Invoked with name=['runc.x86_64'] state=latest lock_timeout=300 conf_file=/tmp/yum.conf allow_downgrade=False autoremove=False bugfix=False disable_gpg_check=False disable_plugin=[] disablerepo=[] download_only=False enable_plugin=[] enablerepo=[] exclude=[] installroot=/ install_repoquery=True install_weak_deps=True security=False skip_broken=False update_cache=False update_only=False validate_certs=True disable_excludes=None download_dir=None list=None releasever=None
Any clarification?
Thanks, Gianluca
BTW the update had a duration of 33 minutes (I selected to not reboot the host): Jun 8, 2021, 10:40:35 AM Host ov300 upgrade was started (User: tekka@mydomain). Jun 8, 2021, 11:13:13 AM Host ov300 upgrade was completed successfully.
At the end if I open a terminal I see:
[root@ov300 ~]# rpm -q ansible ansible-2.9.16-2.el8.noarch [root@ov300 ~]#
and
[root@ov300 ~]# yum update Last metadata expiration check: 0:39:50 ago on Tue 08 Jun 2021 10:41:09 AM CEST. Dependencies resolved.
=============================================================================================================== Package Architecture Version Repository Size
=============================================================================================================== Upgrading: ansible noarch 2.9.21-2.el8 ovirt-4.4-centos-ovirt44 17 M
Transaction Summary
=============================================================================================================== Upgrade 1 Package
Total download size: 17 M Is this ok [y/N]: Operation aborted. [root@ov300 ~]#
Is it expected? Currently in my yum.conf I have:
[main] gpgcheck=1 installonly_limit=3 clean_requirements_on_remove=True best=True skip_if_unavailable=False
Gianluca
_______________________________________________ Users mailing list -- users@ovirt.org To unsubscribe send an email to users-leave@ovirt.org Privacy Statement: https://www.ovirt.org/privacy-policy.html oVirt Code of Conduct: https://www.ovirt.org/community/about/community-guidelines/ List Archives: https://lists.ovirt.org/archives/list/users@ovirt.org/message/4P6YORNJVZG67E...
On Wed, Jun 9, 2021 at 10:41 AM Dana Elfassy <delfassy@redhat.com> wrote:
Hi Gianluca, Upgrade host time depends on the upgrading that needs to be done
For sure. This is always true. In my case part of the update is updating the host from 8.3 to 8.4, so many packages involved. And my observation is that it seems as if the process updates all the packages one by one as subsequent commands, with an unneeded long time in my opinion. Each time it seems it goes through transaction test and so on... Very very suboptimal if it is true and confirmed. From an ansible point of view it seems that it happens something like this - name: ensure a list of packages installed yum: name: "{{ item }}" state: latest loop: "{{ packages }}" instead of - name: ensure a list of packages installed yum: name: "{{ packages }}" state: latest Where {{ packages }} is a pre-computed (probably in a previous step) list of packages to be updated. Gianluca
You are correct. It was implemented this way taking in consideration the output when checking for updates on RHVH hosts. This implementation is indeed going to be changed, so that upgrade will become faster Dana On Wed, Jun 9, 2021 at 2:19 PM Gianluca Cecchi <gianluca.cecchi@gmail.com> wrote:
On Wed, Jun 9, 2021 at 10:41 AM Dana Elfassy <delfassy@redhat.com> wrote:
Hi Gianluca, Upgrade host time depends on the upgrading that needs to be done
For sure. This is always true. In my case part of the update is updating the host from 8.3 to 8.4, so many packages involved. And my observation is that it seems as if the process updates all the packages one by one as subsequent commands, with an unneeded long time in my opinion. Each time it seems it goes through transaction test and so on... Very very suboptimal if it is true and confirmed.
From an ansible point of view it seems that it happens something like this
- name: ensure a list of packages installed yum: name: "{{ item }}" state: latest loop: "{{ packages }}"
instead of
- name: ensure a list of packages installed yum: name: "{{ packages }}" state: latest
Where {{ packages }} is a pre-computed (probably in a previous step) list of packages to be updated.
Gianluca
On Wed, Jun 9, 2021 at 2:26 PM Dana Elfassy <delfassy@redhat.com> wrote:
You are correct. It was implemented this way taking in consideration the output when checking for updates on RHVH hosts. This implementation is indeed going to be changed, so that upgrade will become faster
Dana
Ok, thanks for confirmation. Indeed in case of ovirt-node-ng (or RHV-H for RHV) the update is only one 1Gb sized package containing the new image, but for plain CentOS Stream / CentOS (or RH EL 8 hosts for RHV) it can contain a hundred of packages and so a hundred of transactions vs one. Gianluca
participants (2)
-
Dana Elfassy -
Gianluca Cecchi