
Hi all, I noticed 2 bugs in the code setting up a test: 1. losetup is run with non-existent -D option 2. build does not fail although losetup failed In the current state, test may run with loop devices left from another test - this used to cause random failures for vdsm tests. See log bellow. Nir ---- 22:46:04 Triggered by Gerrit: https://gerrit.ovirt.org/60064 22:46:04 Building remotely on el6-vm08.phx.ovirt.org (el6 vm phx nested) in workspace /home/jenkins/workspace/ovirt-imageio_master_check-patch-fc23-x86_64 22:46:04 Cloning the remote Git repository 22:46:04 Cloning repository git://gerrit.ovirt.org/ovirt-imageio.git 22:46:04 > git init /home/jenkins/workspace/ovirt-imageio_master_check-patch-fc23-x86_64/ovirt-imageio # timeout=10 22:46:04 Fetching upstream changes from git://gerrit.ovirt.org/ovirt-imageio.git 22:46:04 > git --version # timeout=10 22:46:04 > git fetch --tags --progress git://gerrit.ovirt.org/ovirt-imageio.git +refs/heads/*:refs/remotes/origin/* 22:46:06 > git config remote.origin.url git://gerrit.ovirt.org/ovirt-imageio.git # timeout=10 22:46:06 > git config --add remote.origin.fetch +refs/heads/*:refs/remotes/origin/* # timeout=10 22:46:06 > git config remote.origin.url git://gerrit.ovirt.org/ovirt-imageio.git # timeout=10 22:46:06 Cleaning workspace 22:46:06 > git rev-parse --verify HEAD # timeout=10 22:46:06 No valid HEAD. Skipping the resetting 22:46:06 > git clean -fdx # timeout=10 22:46:06 Pruning obsolete local branches 22:46:06 Fetching upstream changes from git://gerrit.ovirt.org/ovirt-imageio.git 22:46:06 > git fetch --tags --progress git://gerrit.ovirt.org/ovirt-imageio.git refs/changes/64/60064/1 --prune 22:46:06 > git rev-parse 19ce0a36c76c3d95658fecd8f59077fc9641e24c^{commit} # timeout=10 22:46:06 Checking out Revision 19ce0a36c76c3d95658fecd8f59077fc9641e24c (master) 22:46:06 > git config core.sparsecheckout # timeout=10 22:46:06 > git checkout -f 19ce0a36c76c3d95658fecd8f59077fc9641e24c 22:46:06 > git rev-parse FETCH_HEAD^{commit} # timeout=10 22:46:06 > git rev-list 77201b6f671a163a60a2a4f4db89d7c9611950d3 # timeout=10 22:46:07 > git branch -a # timeout=10 22:46:07 > git rev-parse remotes/origin/0.2^{commit} # timeout=10 22:46:07 > git rev-parse remotes/origin/master^{commit} # timeout=10 22:46:07 Cloning the remote Git repository 22:46:07 Cloning repository git://gerrit.ovirt.org/jenkins.git 22:46:07 > git init /home/jenkins/workspace/ovirt-imageio_master_check-patch-fc23-x86_64/jenkins # timeout=10 22:46:07 Fetching upstream changes from git://gerrit.ovirt.org/jenkins.git 22:46:07 > git --version # timeout=10 22:46:07 > git fetch --tags --progress git://gerrit.ovirt.org/jenkins.git +refs/heads/*:refs/remotes/origin/* 22:46:09 > git config remote.origin.url git://gerrit.ovirt.org/jenkins.git # timeout=10 22:46:09 > git config --add remote.origin.fetch +refs/heads/*:refs/remotes/origin/* # timeout=10 22:46:09 > git config remote.origin.url git://gerrit.ovirt.org/jenkins.git # timeout=10 22:46:09 Cleaning workspace 22:46:09 > git rev-parse --verify HEAD # timeout=10 22:46:09 No valid HEAD. Skipping the resetting 22:46:09 > git clean -fdx # timeout=10 22:46:09 Pruning obsolete local branches 22:46:09 Fetching upstream changes from git://gerrit.ovirt.org/jenkins.git 22:46:09 > git fetch --tags --progress git://gerrit.ovirt.org/jenkins.git +refs/heads/*:refs/remotes/origin/* --prune 22:46:09 > git rev-parse origin/master^{commit} # timeout=10 22:46:09 Checking out Revision 25d8aa66b047b45ab52bc84ed7dfffedc5a3e222 (origin/master) 22:46:09 > git config core.sparsecheckout # timeout=10 22:46:09 > git checkout -f 25d8aa66b047b45ab52bc84ed7dfffedc5a3e222 22:46:09 > git rev-list 25d8aa66b047b45ab52bc84ed7dfffedc5a3e222 # timeout=10 22:46:09 > git branch -a # timeout=10 22:46:09 > git rev-parse remotes/origin/master^{commit} # timeout=10 22:46:09 [ovirt-imageio_master_check-patch-fc23-x86_64] $ /bin/bash -e /tmp/hudson3544223073540002444.sh 22:46:09 shell-scripts/cleanup_slave.sh 22:46:09 ################################################################### 22:46:09 # Cleaning up slave # 22:46:09 ################################################################### 22:46:09 Filesystem Size Used Avail Use% Mounted on 22:46:09 /dev/vda3 30G 8.0G 21G 29% / 22:46:09 tmpfs 3.9G 0 3.9G 0% /dev/shm 22:46:09 /dev/vda1 93M 31M 58M 35% /boot 22:46:09 --------------------------------------------------------------- 22:46:09 Cleaning up postgres databases 22:46:09 /tmp ~/workspace/ovirt-imageio_master_check-patch-fc23-x86_64 22:46:09 ~/workspace/ovirt-imageio_master_check-patch-fc23-x86_64 22:46:09 done 22:46:09 Cleaning up journal logs (if any) 22:46:09 journald not running, skipping 22:46:09 Cleaning up /var/tmp 22:46:09 done 22:46:09 Emptying some common logs 22:46:09 /var/log/wtmp 22:46:09 Done 22:46:09 /home/jenkins/workspace/ovirt-engine_master_check-patch-el7-x86_64 22:46:10 Making sure there are no device mappings... 22:46:10 Removing the used loop devices... 22:46:10 losetup: invalid option -- 'D' 22:46:10 22:46:10 Usage: 22:46:10 losetup loop_device give info 22:46:10 losetup -a | --all list all used 22:46:10 losetup -d | --detach <loopdev> [<loopdev> ...] delete 22:46:10 losetup -f | --find find unused 22:46:10 losetup -c | --set-capacity <loopdev> resize 22:46:10 losetup -j | --associated <file> [-o <num>] list all associated with <file> 22:46:10 losetup [ options ] {-f|--find|loopdev} <file> setup 22:46:10 22:46:10 Options: 22:46:10 -e | --encryption <type> enable data encryption with specified <name/num> 22:46:10 -h | --help this help 22:46:10 -o | --offset <num> start at offset <num> into file 22:46:10 --sizelimit <num> loop limited to only <num> bytes of the file 22:46:10 -p | --pass-fd <num> read passphrase from file descriptor <num> 22:46:10 -r | --read-only setup read-only loop device 22:46:10 --show print device name (with -f <file>) 22:46:10 -v | --verbose verbose mode 22:46:10 22:46:10 22:46:10 Usage: 22:46:10 losetup loop_device give info 22:46:10 losetup -a | --all list all used 22:46:10 losetup -d | --detach <loopdev> [<loopdev> ...] delete 22:46:10 losetup -f | --find find unused 22:46:10 losetup -c | --set-capacity <loopdev> resize 22:46:10 losetup -j | --associated <file> [-o <num>] list all associated with <file> 22:46:10 losetup [ options ] {-f|--find|loopdev} <file> setup 22:46:10 22:46:10 Options: 22:46:10 -e | --encryption <type> enable data encryption with specified <name/num> 22:46:10 -h | --help this help 22:46:10 -o | --offset <num> start at offset <num> into file 22:46:10 --sizelimit <num> loop limited to only <num> bytes of the file 22:46:10 -p | --pass-fd <num> read passphrase from file descriptor <num> 22:46:10 -r | --read-only setup read-only loop device 22:46:10 --show print device name (with -f <file>) 22:46:10 -v | --verbose verbose mode 22:46:10

בתאריך 1 ביולי 2016 01:54, "Nir Soffer" <nsoffer@redhat.com> כתב:
Hi all,
I noticed 2 bugs in the code setting up a test:
1. losetup is run with non-existent -D option 2. build does not fail although losetup failed
It seems the job ran on an el6 slave. My guess would be that '-D' works on newer OSes (afk so can`t check atm). We should probably just drop the el6 slaves at this point. Mock should probably be enough to keep testing engine on el6 for the 3.6 remaining life time.

Evgheni, can you shut down the el6 slaves and power up fedora or el7 instead? We should have enough powered off to reduce the load. On Jul 1, 2016 9:50 AM, "Barak Korren" <bkorren@redhat.com> wrote:
בתאריך 1 ביולי 2016 01:54, "Nir Soffer" <nsoffer@redhat.com> כתב:
Hi all,
I noticed 2 bugs in the code setting up a test:
1. losetup is run with non-existent -D option 2. build does not fail although losetup failed
It seems the job ran on an el6 slave. My guess would be that '-D' works on newer OSes (afk so can`t check atm). We should probably just drop the el6 slaves at this point. Mock should probably be enough to keep testing engine on el6 for the 3.6 remaining life time.
_______________________________________________ Infra mailing list Infra@ovirt.org http://lists.ovirt.org/mailman/listinfo/infra

Hi everyone, I took offline the three el6 slaves we had on jenkins.ovirt.org: el6-vm08.phx.ovirt.org el6-vm09.phx.ovirt.org el6-vm10.phx.ovirt.org and enabled three el7 hosts instead: el7-vm08.phx.ovirt.org el7-vm09.phx.ovirt.org el7-vm10.phx.ovirt.org In the long run we might need to adjust the job to use labels properly to not end up on el6. Regards, Evgheni Dereveanchin ----- Original Message ----- From: "Eyal Edri" <eedri@redhat.com> To: "Barak Korren" <bkorren@redhat.com>, "Evgheni Dereveanchin" <ederevea@redhat.com> Cc: "infra" <infra@ovirt.org>, "Nir Soffer" <nsoffer@redhat.com> Sent: Friday, 1 July, 2016 9:17:28 AM Subject: Re: 2 bugs in test setup Evgheni, can you shut down the el6 slaves and power up fedora or el7 instead? We should have enough powered off to reduce the load. On Jul 1, 2016 9:50 AM, "Barak Korren" <bkorren@redhat.com> wrote:
בתאריך 1 ביולי 2016 01:54, "Nir Soffer" <nsoffer@redhat.com> כתב:
Hi all,
I noticed 2 bugs in the code setting up a test:
1. losetup is run with non-existent -D option 2. build does not fail although losetup failed
It seems the job ran on an el6 slave. My guess would be that '-D' works on newer OSes (afk so can`t check atm). We should probably just drop the el6 slaves at this point. Mock should probably be enough to keep testing engine on el6 for the 3.6 remaining life time.
_______________________________________________ Infra mailing list Infra@ovirt.org http://lists.ovirt.org/mailman/listinfo/infra

There are still jobs requiring "el6" label so I had to take one of the el6 slaves back online to avoid the build queue from growing. Here's a job that only runs on hosts labelled "el6": http://jenkins.ovirt.org/job/ovirt-engine_3.6_upgrade-db-from-3.5_el6_create... What was the job that was failing? Can we change it to make it avoid el6 hosts? Regards, Evgheni Dereveanchin ----- Original Message ----- From: "Evgheni Dereveanchin" <ederevea@redhat.com> To: "Eyal Edri" <eedri@redhat.com> Cc: "infra" <infra@ovirt.org> Sent: Friday, 1 July, 2016 9:26:06 AM Subject: Re: 2 bugs in test setup Hi everyone, I took offline the three el6 slaves we had on jenkins.ovirt.org: el6-vm08.phx.ovirt.org el6-vm09.phx.ovirt.org el6-vm10.phx.ovirt.org and enabled three el7 hosts instead: el7-vm08.phx.ovirt.org el7-vm09.phx.ovirt.org el7-vm10.phx.ovirt.org In the long run we might need to adjust the job to use labels properly to not end up on el6. Regards, Evgheni Dereveanchin ----- Original Message ----- From: "Eyal Edri" <eedri@redhat.com> To: "Barak Korren" <bkorren@redhat.com>, "Evgheni Dereveanchin" <ederevea@redhat.com> Cc: "infra" <infra@ovirt.org>, "Nir Soffer" <nsoffer@redhat.com> Sent: Friday, 1 July, 2016 9:17:28 AM Subject: Re: 2 bugs in test setup Evgheni, can you shut down the el6 slaves and power up fedora or el7 instead? We should have enough powered off to reduce the load. On Jul 1, 2016 9:50 AM, "Barak Korren" <bkorren@redhat.com> wrote:
בתאריך 1 ביולי 2016 01:54, "Nir Soffer" <nsoffer@redhat.com> כתב:
Hi all,
I noticed 2 bugs in the code setting up a test:
1. losetup is run with non-existent -D option 2. build does not fail although losetup failed
It seems the job ran on an el6 slave. My guess would be that '-D' works on newer OSes (afk so can`t check atm). We should probably just drop the el6 slaves at this point. Mock should probably be enough to keep testing engine on el6 for the 3.6 remaining life time.
_______________________________________________ Infra mailing list Infra@ovirt.org http://lists.ovirt.org/mailman/listinfo/infra
_______________________________________________ Infra mailing list Infra@ovirt.org http://lists.ovirt.org/mailman/listinfo/infra

I think you can disable this job if we can verify this code already runs on standard ci Eli - do we test upgrade scripts in automation/check-patch.sh? anyway, you can disable any 3.5 jobs, so this specific one can be removed. On Mon, Jul 4, 2016 at 3:48 PM, Evgheni Dereveanchin <ederevea@redhat.com> wrote:
There are still jobs requiring "el6" label so I had to take one of the el6 slaves back online to avoid the build queue from growing.
Here's a job that only runs on hosts labelled "el6":
http://jenkins.ovirt.org/job/ovirt-engine_3.6_upgrade-db-from-3.5_el6_create...
What was the job that was failing? Can we change it to make it avoid el6 hosts?
Regards, Evgheni Dereveanchin
----- Original Message ----- From: "Evgheni Dereveanchin" <ederevea@redhat.com> To: "Eyal Edri" <eedri@redhat.com> Cc: "infra" <infra@ovirt.org> Sent: Friday, 1 July, 2016 9:26:06 AM Subject: Re: 2 bugs in test setup
Hi everyone,
I took offline the three el6 slaves we had on jenkins.ovirt.org: el6-vm08.phx.ovirt.org el6-vm09.phx.ovirt.org el6-vm10.phx.ovirt.org
and enabled three el7 hosts instead:
el7-vm08.phx.ovirt.org el7-vm09.phx.ovirt.org el7-vm10.phx.ovirt.org
In the long run we might need to adjust the job to use labels properly to not end up on el6.
Regards, Evgheni Dereveanchin
----- Original Message ----- From: "Eyal Edri" <eedri@redhat.com> To: "Barak Korren" <bkorren@redhat.com>, "Evgheni Dereveanchin" < ederevea@redhat.com> Cc: "infra" <infra@ovirt.org>, "Nir Soffer" <nsoffer@redhat.com> Sent: Friday, 1 July, 2016 9:17:28 AM Subject: Re: 2 bugs in test setup
Evgheni, can you shut down the el6 slaves and power up fedora or el7 instead?
We should have enough powered off to reduce the load. On Jul 1, 2016 9:50 AM, "Barak Korren" <bkorren@redhat.com> wrote:
בתאריך 1 ביולי 2016 01:54, "Nir Soffer" <nsoffer@redhat.com> כתב:
Hi all,
I noticed 2 bugs in the code setting up a test:
1. losetup is run with non-existent -D option 2. build does not fail although losetup failed
It seems the job ran on an el6 slave. My guess would be that '-D' works
on
newer OSes (afk so can`t check atm). We should probably just drop the el6 slaves at this point. Mock should probably be enough to keep testing engine on el6 for the 3.6 remaining life time.
_______________________________________________ Infra mailing list Infra@ovirt.org http://lists.ovirt.org/mailman/listinfo/infra
_______________________________________________ Infra mailing list Infra@ovirt.org http://lists.ovirt.org/mailman/listinfo/infra
-- Eyal Edri Associate Manager RHEV DevOps EMEA ENG Virtualization R&D Red Hat Israel phone: +972-9-7692018 irc: eedri (on #tlv #rhev-dev #rhev-integ)
participants (4)
-
Barak Korren
-
Evgheni Dereveanchin
-
Eyal Edri
-
Nir Soffer