Opened 11 years ago

Closed 10 years ago

#256 closed defect (worksforme)

Mondo restore failure

Reported by: hugu Owned by: bruno
Priority: high Milestone: 2.2.7
Component: mondo Version: 2.2.5
Severity: normal Keywords: restore failure
Cc:

Description

I'd like to restore a mondo NFS backup but it fails with last message :

sh:can't access tty;job control turned off

The platform is : IBM blade HS21 on RHEL5

Versions : mondo-2.2.5-1.rhel5, mindi-2.0.1-1.rhel5, mindi-busybox-1.7.3-1.rhel5

Attachments (7)

mondoarchive.log (158.0 KB) - added by hugu 11 years ago.
mondoarchive.log.log (141.0 KB) - added by hugu 11 years ago.
Mondo-restore-pb-Beginning.jpg (78.8 KB) - added by hugu 11 years ago.
NFS restore: first problem
Mondo-restore-pb.jpg (61.6 KB) - added by hugu 11 years ago.
NFS restore: last screen
PXE-default.txt (308 bytes) - added by hugu 11 years ago.
PXE parameters
default.txt (320 bytes) - added by nbarthelemy 11 years ago.
pxelinux.cfg file
Mondo-restore-pb-20080828.jpg (74.8 KB) - added by hugu 11 years ago.

Download all attachments as: .zip

Change History (24)

comment:1 Changed 11 years ago by bruno

  • Status changed from new to assigned

Problems were fixed on 2.2.5 for NFS support with RHEL5. Could you try with the latest 2.2.6 which was just made today !

TIA, Bruno.

comment:2 Changed 11 years ago by hugu

I get the same error message with mondo 2.2.6 version.

I will try on CD to see if the pb really lies on NFS (I can find some warnings in mondoarchive.log)

Thanks, Hugu

Changed 11 years ago by hugu

comment:3 Changed 11 years ago by bruno

You're still using mindi 2.0.1 and not the latest mindi 2.0.2. THat may explain some of the problems with 2.2.6.

Also what hardware do those lines support: i5000_edac 12481 0 edac_mc 26257 1 i5000_edac

? They are not supported by mindi as of now. Can you also provide your /var/log/mondorestore.log file And a picture of what happens before the one you attached (Shift PgUp? to go back in screen history)

comment:4 Changed 11 years ago by hugu

i5000_edac seems to be for

Intel 5000(P/V/X) class Memory Controllers kernel module

I've upgraded to mindi 2.0.2.

NFS restore : failure (still the same problem)

CD restore : failure with lot of messages "/sbin/init: line xxx: mount: not found" but network is OK. I have no way to get easily mondorestore.log or other screens. It could be useful that busybox provides a simple way (ftp, rcp) to push data elsewhere.

Which way to continue ?

Changed 11 years ago by hugu

Changed 11 years ago by hugu

NFS restore: first problem

Changed 11 years ago by hugu

NFS restore: last screen

comment:5 Changed 11 years ago by hugu

I've just updated attachments after upgrades to mondo 2.2.6 and mindi 2.0.2

Changed 11 years ago by hugu

PXE parameters

comment:6 Changed 11 years ago by bruno

the nfsmount parameter of your PXE setup is incorrect. It should be nfsmount=194.214.218.85:/path/to/the/nfs/mount or nfsmount=194.214.218.109:/export/mondorescue as in the original backup

comment:7 Changed 11 years ago by bruno

Have you had time to confirm tht this is now solved ?

comment:8 Changed 11 years ago by bruno

  • Resolution set to worksforme
  • Status changed from assigned to closed

Changed 11 years ago by nbarthelemy

pxelinux.cfg file

comment:9 follow-up: Changed 11 years ago by nbarthelemy

  • Priority changed from normal to high
  • Resolution worksforme deleted
  • Status changed from closed to reopened

Hi, I seems to have the same issue during the restoration process, after making a successfully backup of Virtual Host based on centos 5 OS.

Note : The NFS files system where iso images are located is correctly mounted by "mindi" during the boot of the PXE restore process.

Here is my configuration :

Linux rhcsmgt 2.6.18-53.el5 #1 SMP

mondo-2.2.6-1.rhel5 mindi-2.0.2-1.rhel5 mindi-busybox-1.7.3-1.rhel5 deps : afio-2.4.7-1.rhel5 buffer-1.19-1.rhel5

Please see my attached pxelinux.cfg config file.

Best Regards,

comment:10 Changed 11 years ago by nbarthelemy

Hi,

After checking the content of initrd generated by mondo it seems to missing some libraries and some binaries comparing to the content of the file /etc/mindi/deplist.txt.

for example, mondorestore and "attached" libraries are missing so the script "post-init" cannot be correctly executed.

So when i add them manually (missing binaries and libaries) in initrd the process can continue.

I don't know why all items (libraries and binaries) listed in /etc/mindi/deplist.txt file are not present in initrd generated by mondo ?

best regards.

comment:11 Changed 11 years ago by bruno

The initrd doesn't contain everything. The rest is taken from the backup media. So the initrd only provides what is needed to mount that media (NIC driver + NFS + ext2/3,xfs,resierfs,... cdrom, ...loop)

As long as you cannot mount that ISO image, your restore is failing.

What's in your /var/log/mondorestore.log ?

comment:12 Changed 11 years ago by nbarthelemy

Hi,

I supposed this because i have the same behavior when i use a USB key as archive media. Altough when i use a DVD as backup media the initrd.img contains mondorestore binary and all dependencies libraries.

/var/log/mondorestore.log file is located on the client machine called mondotest. I can't send it to you, because i can't mount any file system thrue nfs. When i try to mount nfs file system coorectly exported from my server, the client kernel answer "RPC failed to contact portmap" in /var/log/messages file on my mondotest machine. I can see the request on my NFS server log file.

I will try again tomorrow.

Regards

comment:13 follow-up: Changed 11 years ago by hugu

Despite correction on PXE setup suggested by Bruno on 06/17/08, problem still remains. Mondo-restore-pb.20080828.jpg gives the beginning of problems.

Changed 11 years ago by hugu

comment:14 in reply to: ↑ 13 ; follow-up: Changed 11 years ago by bruno

Replying to hugu:

Despite correction on PXE setup suggested by Bruno on 06/17/08, problem still remains. Mondo-restore-pb.20080828.jpg gives the beginning of problems.

Please update first mindi to its latest version 2.0.3 and redo your backup, extract the content to your PXE server and report back again please.

comment:15 in reply to: ↑ 14 Changed 11 years ago by bruno

Replying to bruno:

Please update first mindi to its latest version 2.0.3 and redo your backup, extract the content to your PXE server and report back again please.

Available at ftp://ftp.mondorescue.org/rhel/5/mindi-2.0.3-1.rhel5.i386.rpm and ftp://ftp.mondorescue.org/rhel/5/mindi-2.0.3-1.rhel5.x86_64.rpm

comment:16 in reply to: ↑ 9 Changed 11 years ago by bruno

Replying to nbarthelemy:

Please see my attached pxelinux.cfg config file.

This file is incorrect. Please check the README.pxe file provided with midni to get the correct syntax to use. Cf: http://trac.mondorescue.org/browser/branches/2.2.6/mindi/README.pxe

comment:17 Changed 10 years ago by bruno

  • Resolution set to worksforme
  • Status changed from reopened to closed
Note: See TracTickets for help on using tickets.