14:01:12 <fabiand> #startmeeting oVirt Node Weekly Meeting 14:01:12 <ovirtbot> Meeting started Mon Apr 27 14:01:12 2015 UTC. The chair is fabiand. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:12 <ovirtbot> Useful Commands: #action #agreed #help #info #idea #link #topic. 14:01:21 <fabiand> #chairs rbarry dougsland tlitovsk 14:01:25 <fabiand> #topic Agenda 14:01:31 <fabiand> #info Other Items 14:01:31 * rbarry here 14:01:37 <fabiand> #topic Other Items 14:01:49 <fabiand> Good morning rbarry 14:02:36 * dougsland here 14:02:58 <fabiand> hey dougsland 14:02:59 <rbarry> Morning fa 14:03:10 <fabiand> :) 14:03:15 <fabiand> So, I don#t have a big agenda .. 14:03:19 <fabiand> The biggest item I still see is 14:03:23 <fabiand> #link https://gerrit.ovirt.org/#/c/39304/ 14:03:25 * tlitovsk here 14:03:30 <fabiand> Review rbarry's dbus job 14:03:31 <fabiand> hey tlitovsk 14:04:12 <fabiand> Besides that I do not have items. 14:04:21 <fabiand> rbarry, dougsland tlitovsk -- do you have anything to discuss? 14:04:31 <rbarry> Nothing 14:04:32 <tlitovsk> nope 14:04:35 <fabiand> #action fabiand and tlitovsk to review https://gerrit.ovirt.org/#/c/39304/ 14:04:50 <dougsland> nop 14:05:35 <fabiand> Perfect 14:05:39 <fabiand> Then thanks guys! 14:05:45 <fabiand> tlitovsk, please review that patch as well 14:05:50 <fabiand> If there is nothing else 14:05:51 <fabiand> going once 14:07:33 <fabiand> twice 14:09:39 <YamakasY> karimb: weird huh 16:03:54 <mr_chris> I have some confusion about qcow2 vs raw in ovirt. Is raw the base layer for all new VMs and qcow2 for snapshots? 16:03:54 <mr_chris> I don't see any option to tell a disk image to use qcow2 or raw. It just defaults to raw. 16:04:00 <mr_chris> Yet I'm still able to make snapshots even though it's using a raw disk image. 16:04:10 <mr_chris> I'm even able to make snapshots when using share iscsi volumes. 16:04:13 <mr_chris> How is it doing that? 17:49:19 <loganmc10> I'm doing some research into oVirt and have a question: what happens if the oVirt engine dies? DB corrupt, etc.. 17:49:37 <loganmc10> Is there a way to just create a new Engine and re-add the hosts in 1 by 1 without any VM downtime? 22:04:41 <jflorian-local> Anyone here? 06:45:13 <tlitovsk_> fabiand, network problmes 06:45:21 <fabiand> tlitovsk, right 08:21:06 <alonbl_> fsimonce: hi! 08:21:15 <fsimonce> alonbl_, hi 08:21:34 <alonbl_> fsimonce: the openstack-java sdk was bumped only in fedora22 to 2.1 08:21:36 <alonbl_> fsimonce: the openstack-java sdk was bumped only in fedora22 to 3.1 08:21:47 <alonbl_> fsimonce: so does it mean that engine master cannot work in fc21? 08:22:01 <alonbl_> fsimonce: fc22 is not out yet... right? 08:23:20 <fsimonce> alonbl_, 3.1 introduces several new changes and we wanted to be cautious before releasing it in fedora... there was a thread with sbonazzo where we decided that to begin with we would have placed the rpms in the ovirt repo (for fc21) so that we could have pulled them in case of problems 08:23:50 <alonbl_> fsimonce: so it has not been done, what do we expect devs to do? 08:24:19 <fsimonce> alonbl_, then it has to be done 08:25:11 <fsimonce> alonbl, there were other problems earlier to run on f21 (e.g. java 8) 08:26:12 <alonbl> fsimonce: this can be overridden or use jeap-6.3, no? 08:26:13 <fsimonce> alonbl, if the other problems are all solved and this is the last obstacle... it should have been already fixed because as far as I know for fc21 the packages should have been in our ovirt repo 08:26:26 <alonbl> fsimonce: ok 08:26:30 <alonbl> fsimonce: thanks 08:30:32 <fsimonce> alonbl, http://lists.ovirt.org/pipermail/devel/2015-March/010082.html 08:31:40 <alonbl> fsimonce: sbonazzo|mtg: thanks! they should also be moved to fc21 static repo 08:32:55 <kripper> it seems like the DB update script didn't work and engine is broken 08:35:15 <sbonazzo|mtg> alonbl: fc21 is not supposed to be supported by master / 3.6 fsimonce do you have builds for fc21 ? If you have it, I can add them to the f21 static repo 08:36:42 <kripper> Caused by: org.postgresql.util.PSQLException: The column name gluster_tuned_profile was not found in this ResultSet. 08:36:56 <alonbl> sbonazzo|mtg: fsimonce: I am building locally, btw: fsimonce you the prefix within tarball is %{name}-%{name}-%{version}, not sure how your fc22 spec works :) 08:37:22 <kripper> # cat 03_06_1260_add_tuned_profile_column_to_vds_groups.sql 08:37:22 <kripper> select fn_db_add_column('vds_groups', 'gluster_tuned_profile', 'VARCHAR(50)');[root 08:39:06 <fsimonce> alonbl, I am not seeing %{name}-%{name}-%{version} 08:39:37 <alonbl> fsimonce: strange...: 08:39:42 <alonbl> # tar -tf /root/rpmbuild/SOURCES/openstack-java-sdk-3.1.1.tar.gz | head 08:39:43 <alonbl> openstack-java-sdk-openstack-java-sdk-3.1.1/ 08:39:45 <alonbl> openstack-java-sdk-openstack-java-sdk-3.1.1/.gitignore 08:39:46 <alonbl> openstack-java-sdk-openstack-java-sdk-3.1.1/.travis.yml 08:39:48 <alonbl> openstack-java-sdk-openstack-java-sdk-3.1.1/LICENSE.txt 08:39:55 <fsimonce> alonbl, cat sources 08:39:56 <alonbl> fsimonce: this is the tarball I downloaded from github 08:40:47 <alonbl> fsimonce: different sha1sum, have you created your own tarball? 08:41:02 <fsimonce> -2e798b6f329381a1b95e08f133d0a577 openstack-java-sdk-3.0.6.tar.gz 08:41:02 <fsimonce> +f2d7ac98356b6603c0fa4687d1409cc3 openstack-java-sdk-3.1.1.tar.gz 08:41:36 <alonbl> fsimonce: mine is 4b04c44944674db8cbfd2ebbb368447eb7dbae24, sec, downloading again 08:42:04 <fsimonce> alonbl, tags upstream are in the form "openstack-java-sdk-3.1.1" ...which means that the official tar from github has %{name}-%{name} if you take it from the automated github tar-gzer 08:42:58 <fsimonce> alonbl, you have to use koji where the tar.gz has the proper form 08:42:58 <alonbl> fsimonce: oh! they do not publish proper releases? I selected <Releases> not <Tags> 08:43:09 <fsimonce> alonbl, releases are built from tags 08:43:27 <fsimonce> alonbl, http://koji.fedoraproject.org/koji/taskinfo?taskID=9584795 08:43:42 <alonbl> fsimonce: github added (readded) the ability to store release files. 08:43:44 <fsimonce> alonbl, that's a scratch-build, when it's done I'll do a full build 08:43:50 <alonbl> fsimonce: there is no need for you to recrate the tarball 08:43:59 <alonbl> fsimonce: you can use: https://github.com/woorea/openstack-java-sdk/archive/openstack-java-sdk-3.1.1.tar.gz 08:44:41 <fsimonce> alonbl, I know... but it's the one that you see with the wrong double %{name} 08:45:00 <alonbl> fsimonce: but this is what upstream releases, it is not wrong it is the source 09:03:40 <danken> dcaro: sbonazzo|mtg: can you disable http://jenkins.ovirt.org/job/vdsm_master_install-rpm-sanity-fc22_created/ until it is ported to use dnf? 09:03:51 <danken> I've opened https://fedorahosted.org/ovirt/ticket/317 to track this task 09:04:42 <sbonazzo|mtg> danken: it's working with dnf / yum, error is : 09:34:27 Error: nothing provides libvirt-daemon >= 1.2.14 needed by vdsm-4.17.0-731.gitd7c9c97.fc22.x86_64. 09:06:40 <YamakasY> morning guys! 09:09:52 <bkorren> tee hee hee 09:14:38 <kripper> emesika: Hi Eli, I upgraded engine and missed 4 sql updates. Where is the script to update the DB? 09:25:32 <eedri|mtg> danken, disabled 09:26:22 <YamakasY> anyone knows why a vm with 50 characters does not start with engine-config -g MaxVmNameLengthNonWindows set on 80 ? 09:33:22 <YamakasY> jvandewege: ping pong 09:33:55 <jvandewege> YamakasY: pong ping 09:34:19 <YamakasY> jvandewege: oh my friend, all well there ? 09:34:34 <jvandewege> YamakasY: yes 09:34:43 <YamakasY> jvandewege: that is good to hear! 09:34:59 <YamakasY> jvandewege: new ovirt/virtualization thingies ? 09:35:24 <jvandewege> YamakasY: no, just keeping up with updates 09:35:45 <danken> eedri|mtg: thanks. 09:35:55 <YamakasY> jvandewege: nice, same here... tho I see some weird MaxVmName Lenght issue 09:36:44 <jvandewege> YamakasY: I saw, don't have a solution. Only thing: did you set it for the right cluster version? 09:37:11 <YamakasY> jvandewege: wasn't it overall ? 09:37:25 <YamakasY> that is a good thing tho 09:37:54 <YamakasY> jvandewege: btw, about clusters and so on... we should improve the foreman templating part for ovirt 09:42:29 <jvandewege> YamakasY: what do you want to improve? 09:44:43 <YamakasY> jvandewege: let's say bootorder, the host you want it to start on and set the HA options for it 09:45:15 <YamakasY> It think that is great and saves you login on ovirt after creation 09:46:11 <jvandewege> YamakasY: think I saw a discussion about pxe boot and depending on which version/distro-version of seabios you had it either worked as expected or not. 09:47:29 <YamakasY> jvandewege: erm no that was something else 09:48:03 <YamakasY> jvandewege: it would be nice to set what you want to boot first, I still don't get why disk is first and no pxe... the HA part I think would be great too 09:52:26 <YamakasY> jvandewege: or are you doing twice management still ? like foreman and than ovirt ? 10:07:13 <jvandewege> YamakasY: yes because VM's need sometime an extra cpu or memory and we're not at a stage that we can rebuild from foreman with those extra settings, yet 10:10:55 <YamakasY> jvandewege: erm, foreman is capable of setting CPU's 10:11:10 <YamakasY> jvandewege: just make a profile for it ? 10:37:27 <eedri|lunch> bkp-pto, ping 10:53:10 <bkp> eedri|lunch: Pong 10:53:42 <eedri|lunch> bkp, hey.. just wondering if you got any emails from attaliasn 10:53:48 <eedri|lunch> bkp, i think our lic was approved 10:54:30 <bkp> eedri|lunch: I saw them in my inbox. Have not opened yet, still digging out from taking yesterday off. 10:55:14 <eedri|lunch> bkp, sure 10:56:24 <bkp> eedri|lunch: Just forwarded you the last (approval!) email from them. It looks good to go to me! 10:56:37 <bkp> eedri|lunch: Thanks for following up on that 10:56:41 <eedri|lunch> bkp, excellent, dcaro already logged in 10:56:48 <eedri|lunch> bkp, we'll add an alias for jira.ovirt.org also 10:56:55 <bkp> eedri|lunch: Noted 11:45:28 <sbonazzo> tal: pkliczew: Hi, can you please merge https://gerrit.ovirt.org/40323 ? 11:50:21 <tal> sbonazzo: Done 11:58:05 <sbonazzo> tal thanks 12:34:15 <alonbl> sbonazzo: hi, not sure what the problem is with fc21 and engine, I installed the jboss eap 6.4 and all seems to be working, any reason not to redistribute this one instead of 7.1.1? 12:39:34 <sbonazzo> alonbl: licensing issues, we can't re-distribute EAP 6.4 12:39:55 <sbonazzo> alonbl: not the binary one at least 12:39:57 <alonbl> sbonazzo: strange, why? 12:40:09 <alonbl> sbonazzo: they permit downloading the binaries 12:40:17 <alonbl> sbonazzo: I did not see any terms 12:41:26 <sbonazzo> alonbl, can't remember the full reason, please dig into devel mailing list, there was an open discussion about re-packaging eap 6.4 12:42:04 <alonbl> sbonazzo: worse case we can ask user to do this manually like fetch restriction in gentoo 12:42:27 <alonbl> sbonazzo: this does not mean user cannot use fc21 12:45:54 <sbonazzo> alonbl: ok, btw, on integration side, we're targeting f22 for 3.6 so we're not planning to commit working on F21 in 3.6 time frame 12:48:09 <alonbl> sbonazzo: still if we are to redistribute something, better to redistribute the eap 12:54:33 <sbonazzo> alonbl: did you checked devel list on the reason about not re-distributing eap? 12:55:27 <alonbl> sbonazzo: no, not that important anyway, I thought we can provide better service for users for 3.5 and maybe 3.4 as well. but you guys ties the distro version for some reason. 12:55:45 <alonbl> sbonazzo: not that important to me. 12:56:11 <sbonazzo> alonbl: we added support to el7 in 3.5.1 12:56:29 <sbonazzo> alonbl: and I started trying to get f21 support in 3.5.2 1 month before F21 has been released 12:56:57 <alonbl> sbonazzo: not sure I follow. jboss-7.1.1 has issues with java-8 12:56:58 <sbonazzo> alonbl: I'm not tied to nothing, I just don't have time to support everything 12:57:14 <sbonazzo> alonbl: ok, that's a known issue since december 2014 12:57:15 <alonbl> sbonazzo: the simplest solution is to use jboss-eap and all is working 12:57:25 <alonbl> sbonazzo: not sure why not provide that solution 12:57:29 <sbonazzo> alonbl: I already suggested that and got a nack 12:57:30 <alonbl> sbonazzo: as I wrote not that important to me. 12:57:55 <alonbl> sbonazzo: the question is how you suggested, if there is a problem in redistribute, then user can do this himself 12:58:10 <alonbl> sbonazzo: at least he will get service 12:59:41 <sbonazzo> alonbl: http://lists.ovirt.org/pipermail/devel/2015-February/009878.html 12:59:54 <sbonazzo> alonbl: google may help you on this things 13:00:33 <sbonazzo> alonbl: not sure that common user can get EAP binaries without violating license 13:00:52 <alonbl> sbonazzo: every one can download it 13:01:04 <alonbl> sbonazzo: there was no special approval 13:02:04 <sbonazzo> alonbl: feel free to reopen the discussion 13:02:25 <alonbl> sbonazzo: again, I do not intend to get involve. 13:02:41 <alonbl> sbonazzo: but if this is the only issue, you can present it differently so people can react differently 13:03:18 <alonbl> sbonazzo: "known issue: in order to run ovirt-enigne on fc21, one can download jbeap-6.4, extract it into /opt and use engine-setup --jboss-home=..." 13:03:36 <alonbl> sbonazzo: but this only my philosphoy of how to support customers. 13:05:00 <alonbl> sbonazzo: anyway, I am scared of the jboss-8 and issues that may follow if it is not fully integrated into fedora when we release 13:07:06 <sbonazzo> alonbl: I can then re-assure you, only 3 person replied to the poll using fedora as os for ovirt 13:07:15 <sbonazzo> alonbl: all the others are using centos 13:08:17 <sbonazzo> alonbl: and also if I'm willing to support also those 3 person I'm targeting to the latest available fedora on 3.6 release 13:08:28 <sbonazzo> alonbl: no capacity for supporting more than one 13:08:44 <alonbl> sbonazzo: once again, there is no really an action item 13:08:46 <alonbl> sbonazzo: but ok 13:58:36 <bkp> I figure Rich was slammed, but Gina and Jim? 13:58:45 <bkp> Whoops. 14:30:49 <sbonazzo> mskrivanek: jhernand: evilissimo2 : do we have a SDK / REST api call for forcing a refresh of the guest agent data? 14:31:10 <evilissimo2> sbonazzo: there's no such thing 14:31:15 <evilissimo2> sbonazzo: not even on the vdsm API 14:31:30 <evilissimo2> the only way to get that is to migrate the vm 14:31:44 <sbonazzo> evilissimo2: I mean on the engine side, not on the VM side 14:32:03 <sbonazzo> evilissimo2: suppose I've guest agent version x, reboot, update to version x +1 14:32:16 <sbonazzo> evilissimo2: and I wan't to refresh the data in the engine 14:32:26 <sbonazzo> by reading again the guest agent info 14:33:09 <evilissimo2> well it's currently read every 15 seconds 14:33:36 <sbonazzo> evilissimo2: thanks 14:33:42 <sbonazzo> evilissimo2: so maybe not an issue 14:34:50 <sbonazzo> evilissimo2: fqdn / ip address are refreshed too? 14:35:09 <evilissimo2> when the guest agent starts, it send all its information 14:35:34 <evilissimo2> if not, it's a bug 14:36:12 <sbonazzo> evilissimo2: thanks 15:46:47 <PA_dot> hello, anu recommendation for master domain ?? ( eg. small one, replicated, .... ?? ) or "standard" size with data ? 15:47:31 <PA_dot> and how to change master domain to diff. volume ?? 16:00:04 <PA_dot> hello, anu recommendation for master domain ?? ( eg. small one, replicated, .... ?? ) or "standard" size with data ? 16:00:13 <PA_dot> and how to change master domain to diff. volume ?? 16:04:59 <PA_dot> dobry den , muzu ?? 17:31:50 <PA_dot> anybody for "journal: End of file while reading data: Input/output error " on gluster ( XFS ) , managed by oVirt 17:49:03 <PA_dot> anybody for "journal: End of file while reading data: Input/output error " on gluster ( XFS ) , managed by oVirt 18:14:01 <alonbl> nsoffer: can you help PA_dot? 18:22:33 <nsoffer> alonbl, I don't think so 18:23:01 <nsoffer> PA_dot, please contact glusterfs developers with such issues 18:25:46 <alonbl> nsoffer: who is online? shubhendu? shaharh? 18:26:31 <nsoffer> alonbl, I don't know 18:27:04 <alonbl> nsoffer: email address? something? even I do not know to contact "glusterfs developers"... 18:46:13 <nsoffer> alonbl, I don't know - I never got a reply on redhat #gluster channel 06:08:52 <kripper> hi 06:09:05 <kripper> setMasterDomain() is failing 06:09:15 <kripper> in /usr/share/vdsm/storage/sp.py 06:09:33 <kripper> sbonazzo: I sandro, is /usr/share/vdsm/storage/sp.py yours? 06:17:29 <sbonazzo> kripper: good morning :-) no, it's vdsm code under storage, maybe amureini or nsoffer 06:21:54 <kripper> sbonazzo: good morning. Ok, I found a problem VDSM not correctly detecting all storage pools 06:22:24 <kripper> [ vdsClient -s 0 getConnectedStoragePoolsList ] returns 1 storage pool 06:26:13 <sbonazzo> kripper: I suggest to open a BZ about it 06:47:02 <kripper> sbonazzo: do you know what the 'FilesystemBackend' storage domain is? 06:47:22 <kripper> hosted-engine's SD? 06:49:02 <sbonazzo> kripper: it's a class for accessing SD, it's in the ha package 06:49:48 <sbonazzo> kripper: msivak may help you with that 08:07:56 <mvk> ping jhernand 08:16:48 <kripper> msivak: Hi, I wonder if you can help me with the following issue: I'm getting "StoragePoolMasterNotFound" 08:17:07 <kripper> msivak: thus, I'm not able to create VMs 08:17:27 <kripper> msivak: maybe we can debug it together 08:18:16 <kripper> msivak: vsdClient is listing the storage domain 08:28:48 <mtayer> ykaplan: /etc/ovirt-engine/engine.conf.d/10-setup-database.conf 08:54:55 <kripper> what is the easisest way to install VDSM locally from a git clone to test some pulled patches? 08:56:17 <kripper> make install is not working out of the box or I'm missing some "configure" params? 08:57:20 <ewoud> kripper: does http://www.ovirt.org/Vdsm_Developers help? 08:58:02 <kripper> it only explains how to build a RPM 08:58:17 <kripper> I guess "make install" should work 08:58:25 <kripper> bug I'm missing some configurations 08:58:55 <kripper> OSError: No such file or directory: ${prefix}/etc/libvirt/libvirtd.conf 08:59:10 <kripper> ${prefix} is not replaced 09:04:43 <kripper> make clean did it 09:04:57 <kripper> and a chown 09:08:13 <YamakasY> morning all! 09:10:12 <kripper> morning 12:30:55 <danken> bazulay: hi! 12:31:04 <bazulay> mmucha: https://url.corp.redhat.com/79ee8cd 12:56:56 <atkachuk> hi all ! 13:06:52 <atkachuk> I have a little question : I use oVirt Engine Version: 3.5.1.1-1.el6 after delete host (name for ex. "testA") i cannot add this host second time : Error while executing action: Cannot add Host. The Host name is already in use, please choose a unique name and try again. Smb can help me . Where I must delete record 14:54:56 <wilsonwr> The last weekly sync mentioned challenges integrating with Debian due to an ancient libvirt version; I don't know if it's pertinent, but it looks like Debian Jessie has 1.2.9-9 (was 0.9.12.3-1+deb7u1 in wheezy) 15:29:09 <PA_dot> hello, 15:29:49 <PA_dot> how would I run VM on dedicated hosts ?? ( eg. 2 from 4 ) 16:01:57 <wilsonwr> When setting up a developer environment, are ovirt-host-deploy and otopi installed as root? 19:19:15 <prussell> anyone having live merge success after upgrading to 3.5.2? We're running engine on centos6, all our hypervisors are running 3.5.2 Node el7 and we're getting a failure with error code 52 in the engine logs 09:26:55 <Svedrin> is there any way to make a guest disk have a certain ID_SCSI_SERIAL? normally that value is simply missing, can I set it somehow? 09:51:35 <karimb> Svedrin, if libvirt supports it then you could write a hook 09:51:51 <Svedrin> karimb, how would I do that? :) 10:19:20 <Dw_Sn> hello, no way to enforce any VM to be created with VNC instead of SPICE ? like a default settings / option in the engine ? 11:14:06 <tools> Hey I need help 11:15:04 <tools> I am following this guide http://www.ovirt.org/Quick_Start_Guide#Install_Hosts but when I have installed the ovirt repo and run yum install -y ovirt-engine I get this: No package ovirt0engine available 12:28:34 <danken> dcaro: is glance.ovirt.org temporarily down? 12:29:03 <dcaro> danken: no idea who maintains that 12:30:09 <danken> dcaro: gahh 12:30:27 <danken> masayag: do you have an idea who owns glance.ovirt.org ? 12:30:43 <masayag> danken: Oved does 12:31:15 <danken> dcaro: (I'm surised that we don't have a liveliness test, as we have for gerrit.ovirt.org) 12:31:26 <danken> oved ? ^^^ 12:31:27 <oved> danken, indeed. although haven't touched it in a while, and will be happy to pass ownership to someone else. 12:31:36 <oved> danken, any volunteers? 12:32:54 <dcaro> danken: I was not aware of it's creation or anything, you can open a task on infra track to hand it over to infra team 12:32:59 <dcaro> oved: ^^ 12:33:20 <oved> dcaro, how do I do that? 12:33:33 <dcaro> oved: https://fedorahosted.org/ovirt/ 12:35:22 <danken> oved: shouldn't it be ever-ready? it holds appliance images 12:35:49 <oved> danken, it should. Is it down? 12:37:29 <oved> dcaro|food, done. 12:37:34 <oved> dcaro|food, bon appetite. 12:46:42 <danken> oved: it does not answer pings or http... 12:47:32 <oved> danken, http://glance.ovirt.org:9292/ 13:02:37 <ade> anyone know of ovirt works on F22 yet? 13:03:43 <ade> hmmm, maybe Centos 7,1 is a better bet 13:14:07 <kripper1> hi, how do I git push to my own branch? 13:14:13 <kripper1> on gerrit 13:14:38 <kripper1> I was pushing to github, but I guess I will need to create pull request, don't I? 14:00:32 <dyasny> lvernia, ping 14:01:05 <lvernia> dyasny: pong 14:05:48 <kripper1> morning 14:46:16 <z3d3> hello, is possible to use on dc with 2 nodes shared storage glusterfs when I installed node from ovirt-node-iso-3,5? 15:36:51 <YamakasY> does anyone recognize that the engine is handeling non responsive hosts where actually everything is running fine ? 15:41:28 <kripper1> can someone help me to debug a sanlock deadlock? 16:05:25 <glafouille> <YamakasY> : do you run centos 6? 16:06:01 <glafouille> <YamakasY> I met this issue with centos 6 & 7 + vdsm 4.14.10 16:07:13 <glafouille> <YamakasY> I upgraded to vdsm 4.14.16, I still meet the issue with 6.6, but 7.1 is stabilized. 16:08:13 <glafouille> sorry, I mean upgrade from vdsm 4.16.10 to 4.16.14 17:30:17 <YamakasY> glafouille: I see it in 2 cluster happening, one has C6 and C7 and the other one has only C7 20:49:30 <z3d3> hello, is possible to create glusterfs storage on ovirt-hypervisior ? 08:11:14 <YamakasY> glafouille: around? 08:33:50 <stefanlsd> Hi, im doing a hosted-engine --deploy and during setup it quits with Failed to execute stage 'Misc configuration': [Errno 22] Invalid argument: '/rhev/data-center/mnt/172.16.10.1:_volume1_HostedEngine/0bdd41c3-ccf 7-4c01-beb9-de4f1d2950f9/images/950d4194-9977-4e93-ba82-7f0aea7b84ea' (it seems like from the log its during the sanlock stuff). Any ideas? 14:25:22 <feldspar> Hi. My ovirt-engine install claims all hosts are filtered out because they don't have the necessary networks. When I check the hosts, they do have all the networks they should have, but engine doesn't recognze it. Anyone know how ovirt-engine matches the networks of a host with the networks on record in the database? 14:54:15 <mr_chris> Forgive me but Google has failed me here. I have a guest instance in oVirt that will not shut down. Whether I shut it down from within the guest of forcefully power it off, it keeps automatically restarting. 14:54:25 <mr_chris> It does this regardless of if HA is enabled or not. 14:54:37 <mr_chris> What does this and how do I disable it? 15:59:03 <stefanlsd> Hi, im doing a hosted-engine --deploy and during setup it quits with Failed to execute stage 'Misc configuration': [Errno 22] Invalid argument: '/rhev/data-center/mnt/172.16.10.1:_volume1_HostedEngine/0bdd41c3-ccf 7-4c01-beb9-de4f1d2950f9/images/950d4194-9977-4e93-ba82-7f0aea7b84ea' (it seems like from the log its during the sanlock stuff). Any ideas? 16:41:15 <mr_chris> Forgive me but Google has failed me here. I have a guest instance in oVirt that will not shut down. Whether I shut it down from within the guest of forcefully power it off, it keeps automatically restarting. 16:41:24 <mr_chris> It does this regardless of if HA is enabled or not. What does this and how do I disable it 17:33:04 <feldspar> mr_chris: have you checked the properties of the vm? does it happen on all of your hosts, or only on specific hosts? what did the vdsm log (or engine log) note at the time of shutdown/force restart? 17:38:54 <mr_chris> feldspar, Just that host. I'll check the logs. 17:51:18 <mr_chris> feldspar, And now I can't reproduce the problem. Oh well. Thanks anyways. 17:52:23 <feldspar> mr_chris: you're welcome, though I didn't do anything. :) I've had similar problems, and some seem to stem from engine. 17:54:09 <feldspar> mr_chris: maybe you know something about ovirt logical networks? My engine thinks its hosts are all missing some networks, but refreshing the capabilities doesn't note anything. I can manually check that all networks are there. 17:54:50 <feldspar> still, engine is saying I can't start any VMs because there isn't a single host available with the right networks 17:55:08 <mr_chris> feldspar, I know how to copy the config I was taught. My network foo is pretty weak. 17:55:20 <mr_chris> *network-fu 17:56:07 <mr_chris> You're saying the you can ssh in and see the networks but the engine isn't recognizing it? 17:58:39 <feldspar> mr_chris: yes, that's what I usually do. checking the interfaces reveals everything is there, and up until recently engine didn't think much of it. now it won't let me start anything, or tell me why 17:59:54 <feldspar> mr_chris: what do you run on the vdsm hosts? 18:00:33 <mr_chris> feldspar, Centos 7 node. 18:01:57 <feldspar> mr_chris: mine is on centos 6 rpm. 18:02:24 <mr_chris> feldspar, I can say that one thing we've had no issue with is networking. 18:02:37 <mr_chris> Yet. 18:03:46 <feldspar> good to hear. I've had some trouble getting the network going, mostly that ovirtmgmt would show up to be a vm network even after unchecking that in the Networks tab of engine 18:09:29 <feldspar> mr_chris: do you use more than one network? 18:16:55 <mr_chris> feldspar, We use a few vlans. 18:17:07 <mr_chris> And bonding. If you want I can send you a screenshot of what it looks like to the node. 18:20:20 <feldspar> mr_chris: that'd be great 18:21:26 <mr_chris> feldspar, http://i.imgur.com/6e6yW8y.png 18:24:09 <feldspar> mr_chris: thanks, now I have a better understanding of how that works out 18:24:18 <mr_chris> Glad I could help. 18:24:20 <feldspar> that bond is 3x 1GB? 18:24:37 <mr_chris> 10 Gb/s 18:24:57 <mr_chris> It's running on an m1000e chassis. 18:25:13 <feldspar> nice. 18:26:54 <feldspar> I see your ovirtmgmt network is still marked as 'vm'. Maybe the problems with my ovirt-engine started when I deselected that. 18:37:23 <mr_chris> Maybe. 19:04:09 <mr_chris> A good way to kill a centos 7 node: Try to upgrade it from the engine web interface. 19:05:03 <mr_chris> ISO uploads, it begins to attempt to install it, claims to fail due to loss of SSH connectivity, then is left in a failed state. 19:05:40 <mr_chris> Wound up having to mount the ISO through the DRAC and upgrade that way. 19:09:30 <feldspar> mr_chris: that was all remote? maybe the ssh connectivity wasn't really gone 19:09:51 <mr_chris> Don't know. It's just a mess right now. 19:16:44 <feldspar> The same thing has happened to me with a CentOS 6 install. 19:45:53 <kripper> hi, is "add_lockspace fail result -5" normal? 19:51:04 <feldspar> kripper: don't think so. never even saw that message before, where did you find it? 19:54:21 <kripper> I see it all the time 19:54:26 <kripper> I thought it was normal 19:54:36 <kripper> Where are glusters statedumps? 19:54:39 <kripper> on CentOS? 19:58:01 <kripper> found it 13:34:54 <Atonaby> Hi all. I'm a seasoned admin, not a developer, and not accustomed with the Java world. I try to install ovirt on a Debian test machine (Jessie 8.0, the current stable) and I'm like a fish out of water with jBoss and patternfly. I don't find "easy" step-by-step procedure allowing me to solve the obstacle I encounter. 13:35:21 <Atonaby> Anyone has links or so ? I can provide ssh access to a test machine. 16:28:58 <sudoshred> Good day all 03:05:18 <gotgnu> is there a public mirror site for ovirt downloads that supports rsync and not just http or ftp? 09:23:03 * mvk is away: ` 09:49:29 * mvk is back (gone 00:26:27) 12:18:33 <wakemaster> Hello all, I am in the process of installing ovirt on a OVH dedicated server using the hosted-engine setup. One problem I am having is getting the hosted VM to have internet access. No matter what I try I cannot seem to get it to ping even the gateway. Anyone have any experience or help with this? 13:07:15 <didi> wakemaster: is this nested kvm by any chance? 13:08:22 <wakemaster> @didi yes, it is in kvm 13:08:48 <didi> wakemaster: and you have control over the physical host? 13:09:28 <wakemaster> didi: yea, i have centos 7 installed on the bare metal system 13:10:07 <didi> wakemaster: and you run hosted-engine --deploy on these bare metal systems? Or on VMs inside them? 13:11:17 <wakemaster> didi: on the bare meal system. I am at the stage of the installed where i need to run the engine setup inside the VM. But I can't get network connection 13:11:55 <didi> wakemaster: ok, no idea then :-( For nested kvm (that is, hosts that are actually VMs themselves), you need vdsm-hook-macspoof) 13:12:17 <didi> wakemaster: do you use dhcp? 13:12:26 <didi> wakemaster: does the engine VM get an IP from your dhcp server? 13:12:30 <wakemaster> didi: no i am using static ip addresses 13:12:59 <wakemaster> didi: I think it is something weird with the OVH network and how it works. 13:13:11 <didi> wakemaster: that's probable of course 13:13:24 <didi> wakemaster: did they assign you just a single IP address for your hosts? 13:13:43 <wakemaster> didi: they assigned a single IP, and I bought a secondary one 13:13:58 <didi> wakemaster: did you verify that the secondary one works? 13:14:12 <wakemaster> didi: when I bind it to the bare metal system, yes it doe 13:14:17 <wakemaster> *does 13:14:51 <didi> wakemaster: perhaps they accept only a fixed mac address? 13:15:46 <wakemaster> didi: they do, which I was setting up inside of the VM. I am reinstalling it now and going to set that MAC to the bridge as well and see if that helps 13:16:28 <didi> wakemaster: you might find it easier to debug first with some live cd and virt-install, when this works try continuing with hosted-engine 13:17:26 <wakemaster> didi: alright, I will see if I can manage to try that out 13:17:55 <didi> wakemaster: good luck 13:18:19 <wakemaster> didi: thanks, this project turning into a longer one that expected lol 13:19:44 <didi> wakemaster: hosted-engine --deploy also asks you for the MAC address to use for the VM. Once something works, clean up and try again, and use the mac address that worked for you (and ip address that worked, inside the vm) 13:20:30 <didi> wakemaster: other than that, the vm created by hosted-engine --deploy should behave mostly the same as one created by virt-install 13:21:07 <didi> wakemaster: good luck, please report success/failure on the list if possible (and especially if it was interesting to solve) 13:21:16 <didi> wakemaster: thanks and good luck 13:37:11 <wakemaster> didi: the issue was that during the setup step the MAC address must be set to the given MAC from OVH. Using the MACADDR setting in the VM wasn 13:37:17 <wakemaster> 't overriding it 13:37:54 <didi> wakemaster: 'Using the MACADDR setting in the VM'? Where in the vm? 13:38:24 <didi> wakemaster: in /etc/sysconfig/network-scripts/ifcfg-* ? 13:38:43 <wakemaster> didi: yes 13:38:54 <wakemaster> didi: ifcfg-eth0 13:39:41 <didi> wakemaster: weird, it should have worked imo. Didn't check that recently though. You might want to file a bug against the OS (Centos?)... 13:39:47 <didi> wakemaster: anyway thanks for the report! 13:40:08 <didi> wakemaster: not sure what they do exactly or whther they'll consider it a bug, though 13:40:43 <wakemaster> didi: from all settings you can see it appears to work from inside the VM. I think it is more the OVH network. Something going on there with bridged connections 13:41:20 <didi> wakemaster: sounds so. But how does it work then, when you set on vm level? weird 13:41:47 <wakemaster> didi: it works when you set it at the bare metal system level. Not the VM 13:44:14 <didi> wakemaster: you mean just giving the correct mac address to hosted-egnine --deploy, when asked, was not enough? 13:44:36 <didi> wakemaster: so I don't think HA will work for you 13:44:55 <wakemaster> didi: sorry. That is where you have to set it. When the hosted-engine asks. If you don't set it there. Inside of the VM won't work 13:45:03 <didi> wakemaster: ok 13:45:16 <didi> wakemaster: so it's in vm level, but not inside it 13:45:28 <didi> wakemaster: ugly, but sounds good enough 13:46:07 <wakemaster> didi: yes. Still learning proper wording when it comes to virtualization if I confused you at all 13:46:18 <didi> wakemaster: np at all... 13:46:32 <didi> wakemaster: not sure there is even "proper wording"... 13:47:33 <wakemaster> didi: perfect. Well i will get there eventually to make my point clearer then 07:19:02 <YamakasY> morning 08:18:14 <pdangur> dcaro: https://gerrit.eng.lab.tlv.redhat.com/#/c/16038/ 10:22:02 <nostradamuz> do you also have a slow ovirt mirror.., i am downloading with 120kB/s.. takes ages.. 14:02:08 <ovirtbot> fabiand: Error: Can't start another meeting, one is in progress. 14:02:23 <fabiand> #endmeeting