Re: [Gluster-devel] release-3.7 branch created [Was 3.7.0 update]
On 15 Apr 2015, at 12:54, Ravishankar N wrote: > On 04/15/2015 03:31 PM, Justin Clift wrote: >> On 15 Apr 2015, at 08:09, Ravishankar N wrote: >>> On 04/14/2015 11:57 PM, Vijay Bellur wrote: From here on, we would need patches to be explicitly sent on release-3.7 for the content to be included in a 3.7.x release. Please ensure that you send a backport for release-3.7 after the corresponding patch has been accepted on master. Thanks again to everyone who have helped us in getting here. Look forward to more fun and collaboration as we move towards 3.7.0 GA! >>> For the replication arbiter feature, I'm working on the changes that need >>> to be made on the AFR code. Once it gets merged in master, I will >>> back-port it to 3.7 (I'm targeting to get this done before the GA.). Apart >>> from that I don't think there are new regressions since 3.6 for AFR, so we >>> should be good to go. >> How about this one? >> >> * tests/basic/afr/sparse-file-self-heal.t >> (Wstat: 0 Tests: 64 Failed: 35) >> Failed tests: 1-6, 11, 20-30, 33-34, 36, 41, 50-61, 64 >> Happens in master (Mon 30th March - git commit id >> 3feaf1648528ff39e23748ac9004a77595460c9d) >> (hasn't yet been added to BZ) >> Being investigated by: ? >> >> As per: >> >> https://public.pad.fsfe.org/p/gluster-spurious-failures >> >> ;) > > Just tried the test case a couple of times on my laptop on today's master > and with the head at the above commit ID, passes ever time. :-\ Sure. That's the nature of spurious failures. It's not likely to be trivial to track down... but it *is* important. ;) + Justin -- GlusterFS - http://www.gluster.org An open source, distributed file system scaling to several petabytes, and handling thousands of clients. My personal twitter: twitter.com/realjustinclift ___ Gluster-devel mailing list Gluster-devel@gluster.org http://www.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] release-3.7 branch created [Was 3.7.0 update]
On 04/15/2015 03:31 PM, Justin Clift wrote: On 15 Apr 2015, at 08:09, Ravishankar N wrote: On 04/14/2015 11:57 PM, Vijay Bellur wrote: From here on, we would need patches to be explicitly sent on release-3.7 for the content to be included in a 3.7.x release. Please ensure that you send a backport for release-3.7 after the corresponding patch has been accepted on master. Thanks again to everyone who have helped us in getting here. Look forward to more fun and collaboration as we move towards 3.7.0 GA! For the replication arbiter feature, I'm working on the changes that need to be made on the AFR code. Once it gets merged in master, I will back-port it to 3.7 (I'm targeting to get this done before the GA.). Apart from that I don't think there are new regressions since 3.6 for AFR, so we should be good to go. How about this one? * tests/basic/afr/sparse-file-self-heal.t (Wstat: 0 Tests: 64 Failed: 35) Failed tests: 1-6, 11, 20-30, 33-34, 36, 41, 50-61, 64 Happens in master (Mon 30th March - git commit id 3feaf1648528ff39e23748ac9004a77595460c9d) (hasn't yet been added to BZ) Being investigated by: ? As per: https://public.pad.fsfe.org/p/gluster-spurious-failures ;) Just tried the test case a couple of times on my laptop on today's master and with the head at the above commit ID, passes ever time. :-\ -Ravi + Justin -- GlusterFS - http://www.gluster.org An open source, distributed file system scaling to several petabytes, and handling thousands of clients. My personal twitter: twitter.com/realjustinclift ___ Gluster-devel mailing list Gluster-devel@gluster.org http://www.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] release-3.7 branch created [Was 3.7.0 update]
On 04/15/2015 02:39 PM, Kaushal M wrote: > On Wed, Apr 15, 2015 at 1:20 PM, Emmanuel Dreyfus wrote: >> On Wed, Apr 15, 2015 at 12:58:32PM +0530, Kaushal M wrote: >>> I'm looking at this currently. But I'm having a slightly hard time >>> getting Gluster running on Netbsd. I have a netbsd-6.1.5 installation, >>> but cannot get extended attributes to work and volume creation keeps >>> failing. This is preventing me from checking the issue. >> >> You need to format using -O1 since NetBSD does not yet support extended >> attribute on UFS2 layout, which is the default: >> >> newfs -01 /dev/rwd0e >> mount /dev/rwd0e /mnt >> mkdir -p /mnt/.attribute/user /mnt/.attribute/system >> ymount /mnt >> mount -o extattr /dev/rwd0e /mnt > > Doesn't work on 6.1.5. I'm still getting 'Not supported'. Do I need > NetBSD-7 for this to work? Kaushal, I would suggest that we should also check the archived logs on netbsd for a particular run with mgmt_v3-locks failure and try to analyze. Manu, Could you point out to the archive of such instance? ~Atin > >> >>> Manu, if you help with this? In the meantime I'm setting up a netbsd-7 >>> vm to see if that'll work. >> >> We have a lot of VM at rackspace, though I did not track who is working on >> which one, and therefore I have trouble telling you where to go. Look in >> build.gluster.org for a stopped one in netbsd7_regression, login as jeunkins >> using the same password as on the Linux VM, then use su to become root. >> > > We have a huge queue of builds, so I'd rather not block any of them. > If mgmt_v3-locks.t is failing, you could disable it to allow the tests > to pass. > >> -- >> Emmanuel Dreyfus >> m...@netbsd.org > ___ > Gluster-devel mailing list > Gluster-devel@gluster.org > http://www.gluster.org/mailman/listinfo/gluster-devel > -- ~Atin ___ Gluster-devel mailing list Gluster-devel@gluster.org http://www.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] release-3.7 branch created [Was 3.7.0 update]
On 15 Apr 2015, at 08:09, Ravishankar N wrote: > On 04/14/2015 11:57 PM, Vijay Bellur wrote: >> From here on, we would need patches to be explicitly sent on release-3.7 for >> the content to be included in a 3.7.x release. Please ensure that you send a >> backport for release-3.7 after the corresponding patch has been accepted on >> master. >> >> Thanks again to everyone who have helped us in getting here. Look forward to >> more fun and collaboration as we move towards 3.7.0 GA! >> > > For the replication arbiter feature, I'm working on the changes that need to > be made on the AFR code. Once it gets merged in master, I will back-port it > to 3.7 (I'm targeting to get this done before the GA.). Apart from that I > don't think there are new regressions since 3.6 for AFR, so we should be good > to go. How about this one? * tests/basic/afr/sparse-file-self-heal.t (Wstat: 0 Tests: 64 Failed: 35) Failed tests: 1-6, 11, 20-30, 33-34, 36, 41, 50-61, 64 Happens in master (Mon 30th March - git commit id 3feaf1648528ff39e23748ac9004a77595460c9d) (hasn't yet been added to BZ) Being investigated by: ? As per: https://public.pad.fsfe.org/p/gluster-spurious-failures ;) + Justin -- GlusterFS - http://www.gluster.org An open source, distributed file system scaling to several petabytes, and handling thousands of clients. My personal twitter: twitter.com/realjustinclift ___ Gluster-devel mailing list Gluster-devel@gluster.org http://www.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] release-3.7 branch created [Was 3.7.0 update]
On Wed, Apr 15, 2015 at 3:16 PM, Emmanuel Dreyfus wrote: > On Wed, Apr 15, 2015 at 02:39:05PM +0530, Kaushal M wrote: >> Doesn't work on 6.1.5. I'm still getting 'Not supported'. Do I need >> NetBSD-7 for this to work? > > I think I see the problem now: in netbsd-6 branch the relevant options > are not built by default in the GENERIC kernel. They are in netbsd-7 > GENERIC, but for netbsd-6 a rebuilt is necessary. > > But there is a quick fix since NetBSD hapily mixes newer kernel with older > userland: grab latest netbsd-7 kernel, and boot on it while keeping > your userland as is, it should work. > ftp://ftp.fr.netbsd.org/pub/NetBSD-daily/netbsd-7/201504121610Z/i386/binary/kernel This fixed it. xattrs are now working. I should have an analysis of mgmt_v3-locks.t soon. Thanks Manu. > >> We have a huge queue of builds, so I'd rather not block any of them. > > We have 3 VM unused by jenkins ATM for various reasons: nbslave7[025]. > Use any of them, just check that there is not someone else already > working on it (the w command and its IDLE column help) > >> If mgmt_v3-locks.t is failing, you could disable it to allow the tests >> to pass. > > Done already. > > -- > Emmanuel Dreyfus > m...@netbsd.org ___ Gluster-devel mailing list Gluster-devel@gluster.org http://www.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] release-3.7 branch created [Was 3.7.0 update]
On Wed, Apr 15, 2015 at 02:39:05PM +0530, Kaushal M wrote: > Doesn't work on 6.1.5. I'm still getting 'Not supported'. Do I need > NetBSD-7 for this to work? I think I see the problem now: in netbsd-6 branch the relevant options are not built by default in the GENERIC kernel. They are in netbsd-7 GENERIC, but for netbsd-6 a rebuilt is necessary. But there is a quick fix since NetBSD hapily mixes newer kernel with older userland: grab latest netbsd-7 kernel, and boot on it while keeping your userland as is, it should work. ftp://ftp.fr.netbsd.org/pub/NetBSD-daily/netbsd-7/201504121610Z/i386/binary/kernel > We have a huge queue of builds, so I'd rather not block any of them. We have 3 VM unused by jenkins ATM for various reasons: nbslave7[025]. Use any of them, just check that there is not someone else already working on it (the w command and its IDLE column help) > If mgmt_v3-locks.t is failing, you could disable it to allow the tests > to pass. Done already. -- Emmanuel Dreyfus m...@netbsd.org ___ Gluster-devel mailing list Gluster-devel@gluster.org http://www.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] release-3.7 branch created [Was 3.7.0 update]
On Wed, Apr 15, 2015 at 1:20 PM, Emmanuel Dreyfus wrote: > On Wed, Apr 15, 2015 at 12:58:32PM +0530, Kaushal M wrote: >> I'm looking at this currently. But I'm having a slightly hard time >> getting Gluster running on Netbsd. I have a netbsd-6.1.5 installation, >> but cannot get extended attributes to work and volume creation keeps >> failing. This is preventing me from checking the issue. > > You need to format using -O1 since NetBSD does not yet support extended > attribute on UFS2 layout, which is the default: > > newfs -01 /dev/rwd0e > mount /dev/rwd0e /mnt > mkdir -p /mnt/.attribute/user /mnt/.attribute/system > ymount /mnt > mount -o extattr /dev/rwd0e /mnt Doesn't work on 6.1.5. I'm still getting 'Not supported'. Do I need NetBSD-7 for this to work? > >> Manu, if you help with this? In the meantime I'm setting up a netbsd-7 >> vm to see if that'll work. > > We have a lot of VM at rackspace, though I did not track who is working on > which one, and therefore I have trouble telling you where to go. Look in > build.gluster.org for a stopped one in netbsd7_regression, login as jeunkins > using the same password as on the Linux VM, then use su to become root. > We have a huge queue of builds, so I'd rather not block any of them. If mgmt_v3-locks.t is failing, you could disable it to allow the tests to pass. > -- > Emmanuel Dreyfus > m...@netbsd.org ___ Gluster-devel mailing list Gluster-devel@gluster.org http://www.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] release-3.7 branch created [Was 3.7.0 update]
On Wed, Apr 15, 2015 at 12:58:32PM +0530, Kaushal M wrote: > I'm looking at this currently. But I'm having a slightly hard time > getting Gluster running on Netbsd. I have a netbsd-6.1.5 installation, > but cannot get extended attributes to work and volume creation keeps > failing. This is preventing me from checking the issue. You need to format using -O1 since NetBSD does not yet support extended attribute on UFS2 layout, which is the default: newfs -01 /dev/rwd0e mount /dev/rwd0e /mnt mkdir -p /mnt/.attribute/user /mnt/.attribute/system ymount /mnt mount -o extattr /dev/rwd0e /mnt > Manu, if you help with this? In the meantime I'm setting up a netbsd-7 > vm to see if that'll work. We have a lot of VM at rackspace, though I did not track who is working on which one, and therefore I have trouble telling you where to go. Look in build.gluster.org for a stopped one in netbsd7_regression, login as jeunkins using the same password as on the Linux VM, then use su to become root. -- Emmanuel Dreyfus m...@netbsd.org ___ Gluster-devel mailing list Gluster-devel@gluster.org http://www.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] release-3.7 branch created [Was 3.7.0 update]
I'm looking at this currently. But I'm having a slightly hard time getting Gluster running on Netbsd. I have a netbsd-6.1.5 installation, but cannot get extended attributes to work and volume creation keeps failing. This is preventing me from checking the issue. Manu, if you help with this? In the meantime I'm setting up a netbsd-7 vm to see if that'll work. ~kaushal On Wed, Apr 15, 2015 at 9:44 AM, Emmanuel Dreyfus wrote: > Vijay Bellur wrote: > >> release-3.7 has been branched > > mgmt_v3-locks.t is still broken on both master and release-3.7, right? > > -- > Emmanuel Dreyfus > http://hcpnet.free.fr/pubz > m...@netbsd.org > ___ > Gluster-devel mailing list > Gluster-devel@gluster.org > http://www.gluster.org/mailman/listinfo/gluster-devel ___ Gluster-devel mailing list Gluster-devel@gluster.org http://www.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] release-3.7 branch created [Was 3.7.0 update]
On 04/14/2015 11:57 PM, Vijay Bellur wrote: From here on, we would need patches to be explicitly sent on release-3.7 for the content to be included in a 3.7.x release. Please ensure that you send a backport for release-3.7 after the corresponding patch has been accepted on master. Thanks again to everyone who have helped us in getting here. Look forward to more fun and collaboration as we move towards 3.7.0 GA! For the replication arbiter feature, I'm working on the changes that need to be made on the AFR code. Once it gets merged in master, I will back-port it to 3.7 (I'm targeting to get this done before the GA.). Apart from that I don't think there are new regressions since 3.6 for AFR, so we should be good to go. Thanks, Ravi -Vijay ___ Gluster-devel mailing list Gluster-devel@gluster.org http://www.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] release-3.7 branch created [Was 3.7.0 update]
Vijay Bellur wrote: > release-3.7 has been branched mgmt_v3-locks.t is still broken on both master and release-3.7, right? -- Emmanuel Dreyfus http://hcpnet.free.fr/pubz m...@netbsd.org ___ Gluster-devel mailing list Gluster-devel@gluster.org http://www.gluster.org/mailman/listinfo/gluster-devel
Re: [Gluster-devel] release-3.7 branch created [Was 3.7.0 update]
Vijay Bellur wrote: > From here on, we would need patches to be explicitly sent on > release-3.7 for the content to be included in a 3.7.x release. Please > ensure that you send a backport for release-3.7 after the corresponding > patch has been accepted on master. Would it make sense to have this on release-3.7? http://review.gluster.org/10128 We could use it on release branch to reduce the burden from spurious failures in regressions, while we would not use it on master so that people are pushed to fix them. -- Emmanuel Dreyfus http://hcpnet.free.fr/pubz m...@netbsd.org ___ Gluster-devel mailing list Gluster-devel@gluster.org http://www.gluster.org/mailman/listinfo/gluster-devel
