Re: [Vserver] DRBD and vservers
--- Daniel Hokka Zakrisson <[EMAIL PROTECTED]> wrote: > Martin Fick wrote: > > I am using drbd with vservers and I am running > > into a problem trying to make drbd devices go > > secondary, they report a device busy problem. ... > > I have several vservers and each one has its own > > drbd device so that they can be migrated > > independently from one host to another. The > > problem is that when a vserver is stopped and > > another vserver which was started after the first > > vserver is still running, the drbd device for the > > first device remains busy despite the fact that > > it is not mounted in any namespace anymore. It's > > as if vservers keep a reference to any > > filesystems mounted before they were started even > > if they are not visible within the vserver? Is > > there anyway to fix this? > > Yes, you can enable namespace cleanup. I'm not sure > which version of the > patch is in the current Debian package, it might > need an update, but if > you have a working version, you should be able to > touch > /etc/vservers/.defaults/namespace-cleanup and any > guests you start after > that will not copy all the mounts. OK, this seems to work, the drbd devices are no longer busy. The strange part is that it alaso removes all but the last mount from the current df and mount listings in the host namespace. This has another nasty side effect: when using the FileSystem ocf script with heartbeat, it can no longer unmount any of the hidden filesystems so I am back to the same point, unable to failover gracefully. I guess I could modify the FileSystem script to use /proc/mounts instead of the mount command, but it still seems wrong to not be able to see my filesystems when I use df. Is there a better workaround / fix to this namespace cleanup behavior? > > I am using debian unstable with the debian kernel > > 2.6.16-1-vserver-686, the debian vserver tools > 0.2.6, > > the debian util-vserver 0.30.210.1. > > I assume you mean 0.30.210-10? Yes, sorry aptitude cut off the last 0. Thanks! -Martin __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com ___ Vserver mailing list Vserver@list.linux-vserver.org http://list.linux-vserver.org/mailman/listinfo/vserver
Re: [Vserver] DRBD and vservers
Martin Fick wrote: --- Daniel Hokka Zakrisson <[EMAIL PROTECTED]> wrote: Martin Fick wrote: I am using drbd with vservers and I am running into a problem trying to make drbd devices go secondary, they report a device busy problem. ... I have several vservers and each one has its own drbd device so that they can be migrated independently from one host to another. The problem is that when a vserver is stopped and another vserver which was started after the first vserver is still running, the drbd device for the first device remains busy despite the fact that it is not mounted in any namespace anymore. It's as if vservers keep a reference to any filesystems mounted before they were started even if they are not visible within the vserver? Is there anyway to fix this? Yes, you can enable namespace cleanup. I'm not sure which version of the patch is in the current Debian package, it might need an update, but if you have a working version, you should be able to touch /etc/vservers/.defaults/namespace-cleanup and any guests you start after that will not copy all the mounts. OK, this seems to work, the drbd devices are no longer busy. The strange part is that it alaso removes all but the last mount from the current df and mount listings in the host namespace. This has another nasty side effect: when using the FileSystem ocf script with heartbeat, it can no longer unmount any of the hidden filesystems so I am back to the same point, unable to failover gracefully. I guess I could modify the FileSystem script to use /proc/mounts instead of the mount command, but it still seems wrong to not be able to see my filesystems when I use df. Is there a better workaround / fix to this namespace cleanup behavior? That means the Debian package has the original version of the patch, which is slightly broken in that regard. ;) As a quick fix, you can edit /usr/lib*/util-vserver/vserver.functions:_namespaceCleanup and add -l -n to the umount command. -- Daniel Hokka Zakrisson GPG id: 06723412 GPG fingerprint: A455 4DF3 990A 431F FECA 7947 6136 DDA2 0672 3412 ___ Vserver mailing list Vserver@list.linux-vserver.org http://list.linux-vserver.org/mailman/listinfo/vserver
[Vserver] Vserver-plugin for openQRM released
Dear Vserver community, first thank you for the Vserver-project ! Since Vserver is one of the main-stream virtualization components today i am glad to announce that the openQRM-Team created a Vserver-plugin and released it to the open-source world. ("openQRM" is an open-source, data-center management system. Pls find it at sourceforge) This Vserver-plugin adds some unique features to the openQRM-server like : - fast and user-friendly partitioning/virtualization via Vserver - management of remote Vserver-Hosts - logical abstraction of Servers/Services to allow flexible and dynamic migration from real/physical systems to Vserver-partitions - enhanced network virtualization for Vserver (generic network-device names in the partitions) - fully supports shared-images (one filesystem for many partitions) - ... and more As you all here in this list are most familiar with Vserver any comments and feedback would be more than welcome. hope it is usefull for you, many thanks in advance, Matt -- www.openQRM.org - Keeps your Data-Center Up and Running ___ Vserver mailing list Vserver@list.linux-vserver.org http://list.linux-vserver.org/mailman/listinfo/vserver
[Vserver] nodev with secondary interfaces
The Great Flower Page says nodev "can be used to assign primary interfaces which are created by the host or another vserver.". Are secondary interfaces to be avoided? After returning to the documentation I realised I had unwittingly used nodev with a secondary. I haven't noticed any ill effects in the short time (hours) the vserver has been running. Am I likely to? Jim ___ Vserver mailing list Vserver@list.linux-vserver.org http://list.linux-vserver.org/mailman/listinfo/vserver
Re: [Vserver] Vserver-plugin for openQRM released
yes, even if this is not an planned feature. openQRM provides a mechanism called "easy-migration" which supports monitoring and adminstrating local-installed systems (for easy-migrating them later to netbooting filesystem-images). Combining this "easy-migration" with the Linux-VServer plugin will give you the feature to add/remove/monitor partitiions on an existing Linux-VServer server farm. enjoy, Matt On 7/13/06, ADNET Ghislain <[EMAIL PROTECTED]> wrote: does openQRM can be used to manage a farm of vserver box without the PXE installing thing ? I mean just use the open QRM plugin to add/remove/monitor vservers inside an host on all the vservers farm ? Cordialement, Ghislain ADNET. ___ Vserver mailing list Vserver@list.linux-vserver.org http://list.linux-vserver.org/mailman/listinfo/vserver
Re: [Vserver] DRBD and vservers
--- Daniel Hokka Zakrisson <[EMAIL PROTECTED]> wrote: > >>Martin Fick wrote: > > The strange part is that it alaso removes all but > > the last mount from the current df and mount > > listings in the host namespace. This has another > > nasty side effect: when using the FileSystem ocf > > script with heartbeat, it can no longer unmount > > any of the hidden filesystems so I am back to the > > same point, unable to failover gracefully. > That means the Debian package has the original > version of the patch, > which is slightly broken in that regard. ;) As a > quick fix, you can edit > /usr/lib*/util-vserver/vserver.functions:_namespaceCleanup > and add -l -n to the umount command. Great, that was very helpful, it worked splendidly! You have been very helpful. Finally vservers on drbd failing over gracefully using heartbeat! I am a little surprised that it has been so complicated to get right. I have been fighting with the process for about a week now. I have had to modify the drbd ocf, the Filesystem ocf and the vserver.functions scripts along with creating a vserver ocf script. Am I the only one to ever do this? I know that people are using vservers with drbd, are they not using heartbeat? -Martin __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com ___ Vserver mailing list Vserver@list.linux-vserver.org http://list.linux-vserver.org/mailman/listinfo/vserver
Re: [Vserver] DRBD and vservers
Hi Martin so complicated to get right. I have been fighting with the process for about a week now. I spend some more together with Herbert debugging even issues in kernel code of some filesystems drivers - sometimes using OSS is not only fun (I learned mich about kernel and stuff in this time) a vserver ocf script. Am I the only one to ever do this? I know that people are using vservers with drbd, are they not using heartbeat? Most ppl use only one drbd device for the whole machine - so they never run into such problems perhaps its a god idea too add your knowlegde to the drbd+vserver wikipage :) Oliver -- Diese Nachricht wurde digital unterschrieben oliwel's public key: http://www.oliwel.de/oliwel.crt Basiszertifikat: http://www.ldv.ei.tum.de/page72 smime.p7s Description: S/MIME Cryptographic Signature ___ Vserver mailing list Vserver@list.linux-vserver.org http://list.linux-vserver.org/mailman/listinfo/vserver
[Vserver] guest network interface disappears
We are running latest Gentoo vserver with 8 guest instances on it. Most of them are migrated from a xen environment to vserver and are running fine, but we encountered a strange problem there: 3 of the systems lost their network interface 4 times within a week If I enter the guest in such a case, the "ip" or "ifconfig" commands shows no network interfaces and I can only restart the guest to get it to work again for now. Have anyone an idea how it comes to and how can I avoid it? Thanks in advance Martin ___ Vserver mailing list Vserver@list.linux-vserver.org http://list.linux-vserver.org/mailman/listinfo/vserver
Re: [Vserver] DRBD and vservers
--- Oliver Welter <[EMAIL PROTECTED]> wrote: > Hi Martin > > > so complicated to get right. I have been fighting > > with the process for about a week now. > > I spend some more together with Herbert debugging > even issues in kernel > code of some filesystems drivers - sometimes using > OSS is not only fun > (I learned mich about kernel and stuff in this time) > > > a vserver ocf script. Am I the only one to ever > do > > this? I know that people are using vservers with > > drbd, are they not using heartbeat? > > Most ppl use only one drbd device for the whole > machine - so they never > run into such problems > > perhaps its a god idea too add your knowlegde to the > drbd+vserver > wikipage :) Yes, I will add some stuff, although a lot of it are hacks that I'm not sure are worthy (code quality) of sharing. -Martin __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com ___ Vserver mailing list Vserver@list.linux-vserver.org http://list.linux-vserver.org/mailman/listinfo/vserver
Re: [Vserver] UN - vhashify - ing
Corey Wright wrote: On Wed, 12 Jul 2006 14:53:51 -0700 "Roderick A. Anderson" <[EMAIL PROTECTED]> wrote: Is there a neat trick to un-hashify a guest? find / -type f \ | while read FILE; do cp -av ${FILE} ${FILE}.remove-hashification rm ${FILE} mv ${FILE}.remove-hashification ${FILE} done that's just an example, but should convey the idea well enough. I figured based on some posts from a _long_ time ago it would be copy operation but this is very neat. Will this work from both inside and outside the guest? It a filesystem thing being exploited ( utilized probably sounds better ) by Linux-Vserver? It would probably help me understand better what vhashify is doing ... without going through the code. http://archives.linux-vserver.org/200605/0098.html I remember reading this post. Probably stored it some place and have now forgotten where. http://archives.linux-vserver.org/200605/0228.html I seem to remember this one also. Thanks, Rod -- ___ Vserver mailing list Vserver@list.linux-vserver.org http://list.linux-vserver.org/mailman/listinfo/vserver
Re: [Vserver] totem-video-thumbnailer within vserver?
--- Björn Steinbrink <[EMAIL PROTECTED]> wrote: > On 2006.06.14 21:46:44 -0700, Martin Fick wrote: ... > > Looks like it just can't open eth0, the ip is the > > proper ip for the vserver. > > Yep, the ioctl part virtualizes "too much" away, > this was fixed in > 2.1.1-rc7 IIRC. The relevant patch is here: > http://www.13thfloor.at/~doener/vserver/patches/diff-2.6.16.5-vs2.0.2-rc16-devinet_ioctl.diff > Should apply without any problems. Thanks, I waited for the new debian 2.17.1vserver, it seems to have this patch applied. It works fine now! -Martin __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com ___ Vserver mailing list Vserver@list.linux-vserver.org http://list.linux-vserver.org/mailman/listinfo/vserver
Re: [Vserver] UN - vhashify - ing
On Thu, 13 Jul 2006 09:03:29 -0700 "Roderick A. Anderson" <[EMAIL PROTECTED]> wrote: > Corey Wright wrote: > > On Wed, 12 Jul 2006 14:53:51 -0700 > > "Roderick A. Anderson" <[EMAIL PROTECTED]> wrote: > > > > > >>Is there a neat trick to un-hashify a guest? > > > > > > find / -type f \ > > | while read FILE; do > > cp -av ${FILE} ${FILE}.remove-hashification > > rm ${FILE} > > mv ${FILE}.remove-hashification ${FILE} > > done > > > > that's just an example, but should convey the idea well enough. > > Will this work from both inside and outside the guest? It a filesystem > thing being exploited ( utilized probably sounds better ) by > Linux-Vserver? yes, copying a file, "deleting" the original, and moving (or copying) the copy back, will work in both the context of the host & guests, as it is a mechanism based on the filesystem and not namespaces. the above can be done more selectively/intelligently by insuring a file is immutable and unlinkable using either showattr and/or lsattr. by analyzing my hashified files (confirmed with "ls -i"): if (showattr ${FILE} | grep ^ui. >/dev/null); then echo copy, rm, and mv fi or if (showattr ${FILE} | grep ^i >/dev/null); then echo copy, rm, and mv fi i'm not sure which, if any, is correct to find an immutable-but-unlinkable file created in the process of hashifying as documentation on the subject is scarce and i don't feel like reading/reverse-engineering the source code. but that should give you a good jump start on the subject. corey -- [EMAIL PROTECTED] ___ Vserver mailing list Vserver@list.linux-vserver.org http://list.linux-vserver.org/mailman/listinfo/vserver