Home > Unable To > Unable To Access /mnt/source/images/stage2.img

Unable To Access /mnt/source/images/stage2.img

From: Chris Lumens [PATCH] Rename methodData to stage2Data, since that's more appropriate. It is a rather safe bet that one day somebody will be screwed up by this ruse because a CD with a wrong stage2.img was forgotten in a drive with a Currently, the client systems successfully boot (using PXEboot). I can see on the Alt-F3 log that its trying the new sub-directory. http://globalcryptonews.com/unable-to/unable-to-access-hdlist-file-of-update-source.html

Any ideas? I am now doing a http install using the kernel.org site. but when in the installation process does stage2.img get pulled in and run? RHamel9th November 2007, 11:37 PMWhat are your firewall settings on your NFS box?

From: Chris Lumens [PATCH] Add the --stage2= and --repo= options to anaconda. If the solution does not work for you, open a new bug report. RHamel10th November 2007, 12:21 AMSorry man, **** happens. I did a wireshark trace on the port yesterday and saw that NFS appears to be running normally and that the F8 machine IS connecting successfully to the required NFS directory,

For the hdinstall.c in the patch, if you give "anything" as stage2= then the installer looks for /mnt/isodir/images/stage2.img to loopback mount, Well rather than break everything, I'll just break what I something must have changed between F7 and F8 that alters behavior if nfsdir is specified) There's a list of all supported kernel options here which I'd not been aware about. install.img is transferred only once in rescue mode from CD/DVD. For the record, here's the basic process... 1) You need to get the following services working on the boot server: DHCP server, TFTP, NFS 2) Set up your DHCP server to

Format For Printing -XML -Clone This Bug -Top of page First Last Prev Next This bug is not in your last search results. What if some media would happen to be there? Install fails with NAT even though the server is set up with insecure in /etc/exports so NFS via NAT works for installed systems. https://www.redhat.com/archives/anaconda-devel-list/2008-June/msg00049.html This request is not yet committed for inclusion in a release.

Thanks for the suggestion though. Home | New | Search | [?] | Reports | Requests | Help | NewAccount | Log In [x] | Forgot Password Login: [x] | Report Bugzilla Bug Legal [Date Prev][Date The only reliable way I've found to deal with this is by physically removing the drive I want to preserve prior to installation! skellert10th November 2007, 12:52 AMor IPAPPEND 2 APPEND ksdevice=bootif Hmmm.

If found, we use it to prevent ever having to pull the image over the network. other Try HD install insted of NFS I could, but that won't solve my particular need to PXE install a bunch of computers without CD/DVDs ;-) scottro9th November 2007, 01:12 PMNo, NFS If bridged then it may depend on the DHCP server for your LAN.Actually problem happens in both cases...And both DHCP servers return proper values... (after I've downloaded CentOS VMWare appliance it From: Chris Lumens [PATCH] Reorganize the doLoaderMain state machine to make later work easier.

Put the kickstart file in /tmp and change its perms to 777. (/tmp is already 777, I mean change ks.cfg to 777). this contact form Its only if I PXE load the kernel and ramdisk rather than load off CDROM that it fails. Don't think so. Boot into rescue mode from CD/DVD 2.

Top foxb Posts: 1914 Joined: 2006/04/20 19:03:33 Location: Montreal/QC Re: netinstall DNS (vmware server) Quote Postby foxb » 2010/10/05 01:59:41 Here is it:http://centos.mirror.iweb.ca/5/isos/x86 ... Here's the modified grub config line (only the addition of the HTTP kickstart file).... However, it then begins the install, we know that the place to look is at the kickstart file rather than the NFS share. have a peek here Actual results: # cat /tmp/anaconda.log 14:42:27,548 INFO : kernel command line: initrd=initrd.img rescue BOOT_IMAGE=vmlinuz 14:42:27,548 DEBUG : readNetInfo /tmp/s390net not found, early return 14:42:27,548 INFO : anaconda version 13.21.176 on x86_64

Cheers, Stefan It seems NFS installation may require installation -tree instead of DVD.iso . Does anybody know the filename that the F8 installer goes looking for within the NFS directory??? otherwise, we'll * use files on the first iso image and not be able to umount it */ - if (!strncmp(url, "nfs:", 4)) { + if (!strncmp(url, "nfs:", 4) && !FL_STAGE2(flags))

Are you using NFS?

No PXE involved, and all is fine up through entering the NFS server address and directory. stay tuned. Doing what I just mentioned, putting it in /tmp and giving it 777 perms fixed it. . From: Chris Lumens [PATCH] Add support for the repo= and stage2= options, deprecate method=.

All those warnings are harmless - just the byproduct of the order we do some copying in the loader. If you want to specify an anaconda kickstart file, use the 'ks=nfs' directive to point to the file itself. http://rhn.redhat.com/errata/RHBA-2013-1588.html Note You need to log in before you can comment on or make changes to this bug. Check This Out Running tcpdump on UDP 2049 I see the NFS readdirplus command and a reply from the server listing the ISO file (Fedora-8-i386-DVD.iso).

if we have problems, ask */ char * mountCdromImage(struct installMethod * method, char * location, struct loaderData_s * loaderData, moduleInfoSet modInfo, moduleList modLoaded, moduleDeps * modDepsPtr) { return setupCdrom(location, loaderData, modInfo, From: Chris Lumens To: anaconda-devel-list redhat com Cc: Subject: [PATCH] Support stage2= for NFS installs. Then, we know we have to look at the kickstart file. For past versions of fedora, I've not had to put stage2.img onto the network and the documentation doesn't seem to say this should be necessary....

same problem. Not sure your problems are NFS, but most likely PXE. Status: CLOSED ERRATA Aliases: None Product: Red Hat Enterprise Linux 6 Classification: Red Hat Component: anaconda (Show other bugs) Sub Component: --- Version: 6.3 Hardware: Unspecified Linux Priority unspecified Severity unspecified dirName : "."); return url; } - } - else { + } else { free(path); umount("/mnt/isodir"); flags &= ~LOADER_FLAGS_STAGE2; @@ -269,6 +263,7 @@ int done = 0; char * dir

Both the DVD and rescue CD isos are in the same share. Steps to Reproduce: 1. Top toracat Forum Moderator Posts: 7157 Joined: 2006/09/03 16:37:24 Location: California, US Contact: Contact toracat Website Re: netinstall DNS (vmware server) Quote Postby toracat » 2010/10/01 15:37:12 foxb wrote:It look like Additional info: Problem would seem to be in loader.c, where the call to "findAnacondaCD("/mnt/stage2")" will unmount the CD before returning, if rescue mode is selected.

From: Chris Lumens To: anaconda-devel-list redhat com Cc: Subject: [PATCH] mountStage2 no longer takes a directory argument. Since we've narrowed it down to kickstart, I'd try this. (I had other obscure kickstart problems and fixed it this way.) Temporarily add /tmp to your /etc/exports. Had to use pxe and http instead. It's good to have the stage2 grabbed so we're not leaving that mount point busy later on.

I didn't think permissions were the issue in my case, as everything could access my NFS mounts, but for some reason, the ks.cfg file caused problems.