From: Dan M. <dan...@or...> - 2009-02-03 17:25:30
|
(sorry, I accidentally posted this to -devel the first time) Hello -- I am a Xen developer and I would like to experiment with open-sharedroot on rhel5+ocfs2 on Xen but I am having some network problems and have not been able to download the 3.6G DVD iso. Is there any way I can download just the necessary .rpm files? (i386 only is OK for now) Also, I have ocfs2 v1.4.1-1, not 1.3.9-0.1. Will the specific versions of the rpms listed in the how-to work also with ocfs2 1.4.1 or will I need newer versions of the coomonics rpms? And is the rpm listed in the document (and below) as "el6" a typo? Thanks, Dan List of rpms from the how-to: # comoonics-pythonosfix-py-0.1-1 # comoonics-bootimage-listfiles-1.3-6.el5 # SysVinit-comoonics-2.86-14.atix.1 # comoonics-cluster-py-0.1-12 # comoonics-cdsl-py-0.2-11 # comoonics-bootimage-1.3-32 # comoonics-release-0.1-1 # comoonics-cs-py-0.1-54 # comoonics-bootimage-initscripts-1.3-5.el6 # comoonics-bootimage-extras-ocfs2-0.1-1 # comoonics-bootimage-extras-xen-0.1-3 (Only needed for xen Guest) |
From: Marc G. <gr...@at...> - 2009-02-03 20:27:28
|
Hello, On Tuesday 03 February 2009 18:24:39 Dan Magenheimer wrote: > (sorry, I accidentally posted this to -devel the first time) > > Hello -- > > I am a Xen developer and I would like to experiment with > open-sharedroot on rhel5+ocfs2 on Xen but I am having some > network problems and have not been able to download the > 3.6G DVD iso. Is there any way I can download just the > necessary .rpm files? (i386 only is OK for now) > > Also, I have ocfs2 v1.4.1-1, not 1.3.9-0.1. Will the > specific versions of the rpms listed in the how-to > work also with ocfs2 1.4.1 or will I need newer versions > of the coomonics rpms? Normally it should work. I didn't test it with these versions. On thing that comes into my mind is if you are using ocfs2 via openais (this will not work out of the box) or via the oracle cluster stack (that is supposed to be working). I hopefully will release in the next days a new version of the comoonics rpms that are also validated against RHEL5 with ocfs. > > And is the rpm listed in the document (and below) > as "el6" a typo? That's a typo. > > Thanks, > Dan You're welcome. Let us know about your success. Regards Marc. > > List of rpms from the how-to: > # comoonics-pythonosfix-py-0.1-1 > # comoonics-bootimage-listfiles-1.3-6.el5 > # SysVinit-comoonics-2.86-14.atix.1 > # comoonics-cluster-py-0.1-12 > # comoonics-cdsl-py-0.2-11 > # comoonics-bootimage-1.3-32 > # comoonics-release-0.1-1 > # comoonics-cs-py-0.1-54 > # comoonics-bootimage-initscripts-1.3-5.el6 > # comoonics-bootimage-extras-ocfs2-0.1-1 > # comoonics-bootimage-extras-xen-0.1-3 (Only needed for xen Guest) > > --------------------------------------------------------------------------- >--- Create and Deploy Rich Internet Apps outside the browser with > Adobe(R)AIR(TM) software. With Adobe AIR, Ajax developers can use existing > skills and code to build responsive, highly engaging applications that > combine the power of local resources and data with the reach of the web. > Download the Adobe AIR SDK and Ajax docs to start building applications > today-http://p.sf.net/sfu/adobe-com > _______________________________________________ > Open-sharedroot-users mailing list > Ope...@li... > https://lists.sourceforge.net/lists/listinfo/open-sharedroot-users -- Gruss / Regards, Marc Grimme http://www.atix.de/ http://www.open-sharedroot.org/ |
From: Dan M. <dan...@or...> - 2009-02-07 00:43:36
|
Finally got the rpms and installed. One other typo on the rhel5/ocfs2 webpage: comoonics-bootimage-1.3-33 (not -32) as comoonics-bootimage-extras-ocfs2-0.1-1 depends on -33. On to the next step! :-) Do I need a redhat cluster.conf file other than to have the ocfs2 cluster.conf file created from it? In other words, can I just start with my own /etc/ocfs2/cluster.conf file? Or will the other installation steps require /etc/cluster/cluster.conf? Thanks, Dan > -----Original Message----- > From: Marc Grimme [mailto:gr...@at...] > Sent: Tuesday, February 03, 2009 1:27 PM > To: ope...@li... > Cc: Dan Magenheimer > Subject: Re: [OSR-users] rpms for ocfs2+rhel5 > > > Hello, > On Tuesday 03 February 2009 18:24:39 Dan Magenheimer wrote: > > (sorry, I accidentally posted this to -devel the first time) > > > > Hello -- > > > > I am a Xen developer and I would like to experiment with > > open-sharedroot on rhel5+ocfs2 on Xen but I am having some > > network problems and have not been able to download the > > 3.6G DVD iso. Is there any way I can download just the > > necessary .rpm files? (i386 only is OK for now) > > > > Also, I have ocfs2 v1.4.1-1, not 1.3.9-0.1. Will the > > specific versions of the rpms listed in the how-to > > work also with ocfs2 1.4.1 or will I need newer versions > > of the coomonics rpms? > Normally it should work. I didn't test it with these > versions. On thing that > comes into my mind is if you are using ocfs2 via openais > (this will not work > out of the box) or via the oracle cluster stack (that is > supposed to be > working). > > I hopefully will release in the next days a new version of > the comoonics rpms > that are also validated against RHEL5 with ocfs. > > > > And is the rpm listed in the document (and below) > > as "el6" a typo? > That's a typo. > > > > Thanks, > > Dan > You're welcome. > Let us know about your success. > > Regards Marc. > > > > List of rpms from the how-to: > > # comoonics-pythonosfix-py-0.1-1 > > # comoonics-bootimage-listfiles-1.3-6.el5 > > # SysVinit-comoonics-2.86-14.atix.1 > > # comoonics-cluster-py-0.1-12 > > # comoonics-cdsl-py-0.2-11 > > # comoonics-bootimage-1.3-32 > > # comoonics-release-0.1-1 > > # comoonics-cs-py-0.1-54 > > # comoonics-bootimage-initscripts-1.3-5.el6 > > # comoonics-bootimage-extras-ocfs2-0.1-1 > > # comoonics-bootimage-extras-xen-0.1-3 (Only needed for xen Guest) > > > > > -------------------------------------------------------------- > ------------- > >--- Create and Deploy Rich Internet Apps outside the browser with > > Adobe(R)AIR(TM) software. With Adobe AIR, Ajax developers > can use existing > > skills and code to build responsive, highly engaging > applications that > > combine the power of local resources and data with the > reach of the web. > > Download the Adobe AIR SDK and Ajax docs to start building > applications > > today-http://p.sf.net/sfu/adobe-com > > _______________________________________________ > > Open-sharedroot-users mailing list > > Ope...@li... > > https://lists.sourceforge.net/lists/listinfo/open-sharedroot-users > > > > -- > Gruss / Regards, > > Marc Grimme > http://www.atix.de/ http://www.open-sharedroot.org/ > > |
From: Marc G. <gr...@at...> - 2009-02-07 10:13:17
|
On Saturday 07 February 2009 01:42:42 Dan Magenheimer wrote: > Finally got the rpms and installed. One other typo on > the rhel5/ocfs2 webpage: > > comoonics-bootimage-1.3-33 (not -32) > > as comoonics-bootimage-extras-ocfs2-0.1-1 depends > on -33. Thanks. I fixed both typos. You might want to check if those are ok now. > > On to the next step! :-) > > Do I need a redhat cluster.conf file other than to > have the ocfs2 cluster.conf file created from it? > In other words, can I just start with my own > /etc/ocfs2/cluster.conf file? Or will the other > installation steps require /etc/cluster/cluster.conf? Yes, this is something I don't like either but as I had no customer willing to pay for it I took the simplest way ;-). That means up to now you need a /etc/cluster/cluster.conf based on redhat cluster. And on the other side it wasn't possible to extend the ocfs cluster.conf with attributes. So in the current state you will have to create a cluster.conf. This will be the base for the /etc/ocfs2/cluster.conf which will be automatically generated by com-queryclusterconf convert ocfs2. To validate the output you can use the command. It should return something like below: [root@generix3 ~]# com-queryclusterconf convert ocfs2 node: ip_port = 7777 ip_address = 10.10.10.83 number = 3 name = generix3.local cluster = clu_generix node: ip_port = 7777 ip_address = 10.10.10.82 number = 2 name = generix2.local cluster = clu_generix node: ip_port = 7777 ip_address = 10.10.10.84 number = 4 name = generix4.local cluster = clu_generix cluster: node_count = 3 name = clu_generix Does that help and make sense? Marc. -- Gruss / Regards, Marc Grimme http://www.atix.de/ http://www.open-sharedroot.org/ |
From: Dan M. <dan...@or...> - 2009-02-09 23:11:48
|
> Does that help and make sense? Thanks, yes. But I would like to test the process by creating and booting a one-node (or at most two-node) cluster and then "grow" it later to a larger cluster. Is this possible? If so, what parts/scripts of the howto will I have to re-do when I am ready to grow it from one-node (or two-nodes) to many? Thanks, Dan > -----Original Message----- > From: Marc Grimme [mailto:gr...@at...] > Sent: Saturday, February 07, 2009 3:13 AM > To: ope...@li... > Cc: Dan Magenheimer > Subject: Re: [OSR-users] rpms for ocfs2+rhel5 > > > On Saturday 07 February 2009 01:42:42 Dan Magenheimer wrote: > > Finally got the rpms and installed. One other typo on > > the rhel5/ocfs2 webpage: > > > > comoonics-bootimage-1.3-33 (not -32) > > > > as comoonics-bootimage-extras-ocfs2-0.1-1 depends > > on -33. > Thanks. I fixed both typos. You might want to check if those > are ok now. > > > > On to the next step! :-) > > > > Do I need a redhat cluster.conf file other than to > > have the ocfs2 cluster.conf file created from it? > > In other words, can I just start with my own > > /etc/ocfs2/cluster.conf file? Or will the other > > installation steps require /etc/cluster/cluster.conf? > Yes, this is something I don't like either but as I had no > customer willing to > pay for it I took the simplest way ;-). > > That means up to now you need a /etc/cluster/cluster.conf > based on redhat > cluster. And on the other side it wasn't possible to extend the ocfs > cluster.conf with attributes. > > So in the current state you will have to create a > cluster.conf. This will be > the base for the /etc/ocfs2/cluster.conf which will be automatically > generated by com-queryclusterconf convert ocfs2. To validate > the output you > can use the command. It should return something like below: > > [root@generix3 ~]# com-queryclusterconf convert ocfs2 > > node: > ip_port = 7777 > ip_address = 10.10.10.83 > number = 3 > name = generix3.local > cluster = clu_generix > > node: > ip_port = 7777 > ip_address = 10.10.10.82 > number = 2 > name = generix2.local > cluster = clu_generix > > node: > ip_port = 7777 > ip_address = 10.10.10.84 > number = 4 > name = generix4.local > cluster = clu_generix > > cluster: > node_count = 3 > name = clu_generix > > Does that help and make sense? > > Marc. > > -- > Gruss / Regards, > > Marc Grimme > http://www.atix.de/ http://www.open-sharedroot.org/ > > |
From: Dan M. <dan...@or...> - 2009-02-09 23:36:24
|
OK, I've finished the howto, trying to reasonably substitute where necessary for the fact that I am going to boot the osr as a xen guest (running 2.6.28 as the guest kernel). Note that I am running the howto in a Xen EL5 guest booted with 2.6.28 and with an ocfs2 filesystem mounted in the guest as a second disk "/extra"... so wherever I see /mnt/newroot, I am substituting /extra. All seems to go smoothly until I get to the last step... the mkinitrd complains loudly and lengthy. It can't find a list of rpms and skips: - comoonics-cs-xml - kmod-gfs - gfs-utils - cman - openais - lvm2-cluster - OpenIPMI-tools - xen - xen-libs - libvirt Then the /boot disk fills and I get cpio errors. It appears that an intird...img.tmp file is created that is 78MB in size. Oh dear... it also looks like the grub part of the howto has changed /boot/grub on my running system, not on the /extra disk?! And now I can't boot that guest anymore at all! When I have a chance, I will restart with a fresh EL5u2 guest and document every step from the beginning, but if you see any obvious solutions, please let me know. Thanks, Dan > -----Original Message----- > From: Dan Magenheimer > Sent: Monday, February 09, 2009 3:11 PM > To: Marc Grimme; ope...@li... > Subject: Re: [OSR-users] rpms for ocfs2+rhel5 > > > > Does that help and make sense? > > Thanks, yes. But I would like to test the process by > creating and booting a one-node (or at most two-node) > cluster and then "grow" it later to a larger cluster. > Is this possible? If so, what parts/scripts of the > howto will I have to re-do when I am ready to grow > it from one-node (or two-nodes) to many? > > Thanks, > Dan > > > -----Original Message----- > > From: Marc Grimme [mailto:gr...@at...] > > Sent: Saturday, February 07, 2009 3:13 AM > > To: ope...@li... > > Cc: Dan Magenheimer > > Subject: Re: [OSR-users] rpms for ocfs2+rhel5 > > > > > > On Saturday 07 February 2009 01:42:42 Dan Magenheimer wrote: > > > Finally got the rpms and installed. One other typo on > > > the rhel5/ocfs2 webpage: > > > > > > comoonics-bootimage-1.3-33 (not -32) > > > > > > as comoonics-bootimage-extras-ocfs2-0.1-1 depends > > > on -33. > > Thanks. I fixed both typos. You might want to check if those > > are ok now. > > > > > > On to the next step! :-) > > > > > > Do I need a redhat cluster.conf file other than to > > > have the ocfs2 cluster.conf file created from it? > > > In other words, can I just start with my own > > > /etc/ocfs2/cluster.conf file? Or will the other > > > installation steps require /etc/cluster/cluster.conf? > > Yes, this is something I don't like either but as I had no > > customer willing to > > pay for it I took the simplest way ;-). > > > > That means up to now you need a /etc/cluster/cluster.conf > > based on redhat > > cluster. And on the other side it wasn't possible to extend > the ocfs > > cluster.conf with attributes. > > > > So in the current state you will have to create a > > cluster.conf. This will be > > the base for the /etc/ocfs2/cluster.conf which will be > automatically > > generated by com-queryclusterconf convert ocfs2. To validate > > the output you > > can use the command. It should return something like below: > > > > [root@generix3 ~]# com-queryclusterconf convert ocfs2 > > > > node: > > ip_port = 7777 > > ip_address = 10.10.10.83 > > number = 3 > > name = generix3.local > > cluster = clu_generix > > > > node: > > ip_port = 7777 > > ip_address = 10.10.10.82 > > number = 2 > > name = generix2.local > > cluster = clu_generix > > > > node: > > ip_port = 7777 > > ip_address = 10.10.10.84 > > number = 4 > > name = generix4.local > > cluster = clu_generix > > > > cluster: > > node_count = 3 > > name = clu_generix > > > > Does that help and make sense? > > > > Marc. > > > > -- > > Gruss / Regards, > > > > Marc Grimme > > http://www.atix.de/ http://www.open-sharedroot.org/ > > > > > > -------------------------------------------------------------- > ---------------- > Create and Deploy Rich Internet Apps outside the browser with > Adobe(R)AIR(TM) > software. With Adobe AIR, Ajax developers can use existing > skills and code to > build responsive, highly engaging applications that combine > the power of local > resources and data with the reach of the web. Download the > Adobe AIR SDK and > Ajax docs to start building applications > today-http://p.sf.net/sfu/adobe-com > _______________________________________________ > Open-sharedroot-users mailing list > Ope...@li... > https://lists.sourceforge.net/lists/listinfo/open-sharedroot-users > |
From: Marc G. <gr...@at...> - 2009-02-10 06:18:36
|
On Tuesday 10 February 2009 00:35:26 Dan Magenheimer wrote: > OK, I've finished the howto, trying to reasonably substitute > where necessary for the fact that I am going to boot the > osr as a xen guest (running 2.6.28 as the guest kernel). > > Note that I am running the howto in a Xen EL5 guest booted > with 2.6.28 and with an ocfs2 filesystem mounted in the > guest as a second disk "/extra"... so wherever I see > /mnt/newroot, I am substituting /extra. That should not be a problem. BTW. I also made the howto for nodes running in XEN with EL5. So that should fit. > > All seems to go smoothly until I get to the last step... > the mkinitrd complains loudly and lengthy. It can't find > a list of rpms and skips: > - comoonics-cs-xml > - kmod-gfs > - gfs-utils > - cman > - openais > - lvm2-cluster > - OpenIPMI-tools > - xen > - xen-libs > - libvirt You can ignore those. These are only warnings (I don't like them either and they will go away in the future) for rpms not to be found. Most are related to GFS and RedHat Cluster or xen Dom0. You can siltently ignore then. > > Then the /boot disk fills and I get cpio errors. It appears > that an intird...img.tmp file is created that is 78MB in size. Yes. It might be that the boot disk is too small. A workaround would be to first create the initrd in /tmp. I.e mkinitrd /tmp/initrd_sr... and then copy/move it to boot. I've done this whenever the bootdisk was too small. > > Oh dear... it also looks like the grub part of the howto has > changed /boot/grub on my running system, not on the /extra > disk?! And now I can't boot that guest anymore at all! Yes, the bootdisk normally resides on a special disk and therefore must be handled specially. See above for a workaround. > > When I have a chance, I will restart with a fresh EL5u2 > guest and document every step from the beginning, but > if you see any obvious solutions, please let me know. If possible let the /boot be 1G. Regards Marc. -- Gruss / Regards, Marc Grimme http://www.atix.de/ http://www.open-sharedroot.org/ |
From: Marc G. <gr...@at...> - 2009-02-10 06:13:06
|
On Monday 09 February 2009 23:11:18 Dan Magenheimer wrote: > > Does that help and make sense? > > Thanks, yes. But I would like to test the process by > creating and booting a one-node (or at most two-node) > cluster and then "grow" it later to a larger cluster. > Is this possible? If so, what parts/scripts of the > howto will I have to re-do when I am ready to grow > it from one-node (or two-nodes) to many? If ocfs2 does support it we do as well. I think you just need to add a node to the /etc/cluster/cluster.conf build a new initrd and then boot the new node. Additionally you need to have the hostdependent files ready for that node as well. Those will be found at /cluster/cdsl/<nodeid>. Best if this directory is not there it to either copy /cluster/cdsl/default to /cluster/cdsl/<nodeid> and then adapt the hostdep. files. Hope that helps Regrads Marc. -- Gruss / Regards, Marc Grimme http://www.atix.de/ http://www.open-sharedroot.org/ |