Uploading to wrong storage group

[Sebastian Roth]

@dolf Would be great to hear if your issue is solved after upgrading.

Moving this to bug reports…

[Wayne Workman]

Also - I’ve added code to ensure that FTP credentials are set exactly correct during the fog installation now - on both normal and storage installations. Updating will rectify any FTP credentials issues you may have.

[dolf]

Ok so in the mean time I messed up my test PC, so I have to deploy to it before I can try to capture again. When deploying, I got a similar issue, where the image stored on /mnt/olifant/images was “not found”.

So I upgraded to trunk. On build 8499 now.

Now I get something else:

Presumably it is trying to mount the NFS share at /mnt/olifant/images on my server to /images on the client?

[dolf]

After checking permissions, etc. according to the wiki https://wiki.fogproject.org/wiki/index.php?title=Troubleshoot_NFS , I noticed that the last two lines of /etc/exports were missing. This is how it looks now:

/images *(ro,sync,no_wdelay,no_subtree_check,insecure_locks,no_root_squash,insecure,fsid=0)
/images/dev *(rw,async,no_wdelay,no_subtree_check,no_root_squash,insecure,fsid=1)
/mnt/olifant/images *(ro,sync,no_wdelay,no_subtree_check,insecure_locks,no_root_squash,insecure,fsid=0)
/mnt/olifant/images/dev *(rw,async,no_wdelay,no_subtree_check,no_root_squash,insecure,fsid=1)

So both storage groups are on the same machine.

When deploying, the error now changes to “Unable to locate image store”.

I checked and re-submitted the settings of the image and the node on the web interface, just in case. I also checked the presence of the files at /mnt/olifant/images/IMAGE_NAME on the server. Then I cancelled the task and tried to deploy again. The web interface now shows this error:

Failed to create deployment tasking for the following hosts
No Storage Node found for this image

Resubmitting the settings pages for the storage nodes has no effect.

[Wayne Workman]

@dolf Look at the image’s storage groups - make sure the right group is present and set as primary (the primary is a green check box), and remove the other group, and try again.

[dolf]

@Wayne-Workman My image and storage group settings still look exactly like in the first post. I don’t see a checkbox called “primary” anywhere.

With regards to removing a storage group… I have lots of images in both groups, and I need both, so I can’t delete them, lest I lose my images.

[Wayne Workman]

@dolf I’m specifically talking about the images storage group. You’ve not posted a picture of that. Click the image, on the left is storage groups.

[dolf]

Sorry, I wasn’t aware that an image can belong to more than one storage group. The relevant images were in one group (olifant), but it wasn’t marked as primary. It is marked primary now. The same error shows up when trying to deploy:

Note that I did not sensor anything in that image. There are no hosts listed…

[Wayne Workman]

@dolf it could be that the image uploaded to /images instead of the correct directory. Can you check? ls -lahRt /images

[dolf]

Files are in the right places, IMO:

$ tree /mnt/olifant/images/
/mnt/olifant/images/
├── 20160127Temp
│   ├── d1.fixed_size_partitions
│   ├── d1.original.fstypes
│   ├── d1.original.partitions
│   ├── d1.original.swapuuids
│   ├── rec.img.000
│   └── sys.img.000
├── 20160706W7PreSysprep
│   ├── d1.fixed_size_partitions
│   ├── d1.mbr
│   ├── d1.minimum.partitions
│   ├── d1.original.fstypes
│   ├── d1.original.swapuuids
│   ├── d1p1.img
│   ├── d1p2.img
│   └── d1.partitions
├── 20160707W7PreSysprep
│   ├── d1.fixed_size_partitions
│   ├── d1.mbr
│   ├── d1.minimum.partitions
│   ├── d1.original.fstypes
│   ├── d1.original.swapuuids
│   ├── d1p1.img
│   ├── d1p2.img
│   └── d1.partitions
├── box20150731
│   ├── d1.fixed_size_partitions
│   ├── d1.original.fstypes
│   ├── d1.original.partitions
│   ├── d1.original.swapuuids
│   ├── rec.img.000
│   └── sys.img.000
├── dev
├── lab2win720150409
│   ├── d1.fixed_size_partitions
│   ├── d1.original.fstypes
│   ├── d1.original.partitions
│   ├── d1.original.swapuuids
│   ├── rec.img.000
│   └── sys.img.000
├── lab2win720150410
│   ├── d1.fixed_size_partitions
│   ├── d1.original.fstypes
│   ├── d1.original.partitions
│   ├── d1.original.swapuuids
│   ├── rec.img.000
│   └── sys.img.000
├── lab420150716
│   ├── d1.fixed_size_partitions
│   ├── d1.original.fstypes
│   ├── d1.original.partitions
│   ├── d1.original.swapuuids
│   ├── rec.img.000
│   └── sys.img.000
├── thin20150801
│   ├── d1.fixed_size_partitions
│   ├── d1.original.fstypes
│   ├── d1.original.partitions
│   ├── d1.original.swapuuids
│   ├── rec.img.000
│   └── sys.img.000
└── Win7noDriversnoDefaultUser
    ├── rec.img.000
    └── sys.img.000

$ tree /images
/images
├── 20160128W7PreSysprep
│   ├── d1.fixed_size_partitions
│   ├── d1.original.fstypes
│   ├── d1.original.partitions
│   ├── d1.original.swapuuids
│   ├── rec.img.000
│   └── sys.img.000
├── 20160129W7PostSysprep
│   ├── d1.fixed_size_partitions
│   ├── d1.original.fstypes
│   ├── d1.original.partitions
│   ├── d1.original.swapuuids
│   ├── rec.img.000
│   └── sys.img.000
├── dev
├── postdownloadscripts
│   └── fog.postdownload
└── Windows7Software
    ├── rec.img.000
    └── sys.img.000

[Wayne Workman]

@dolf in your first post, you said you manually moved the image. Tree doesn’t show the permissions, which could be suspect. Please check those.

[dolf]

It actually does, using color, which obviously didn’t survive the copy paste. Here is the output with permissions (I already did chmod -R 777 * )

$ ls -lha /images
total 28K
drwxrwxrwx  7 root root 4,0K Jul  9 10:57 .
drwxr-xr-x 28 root root 4,0K Jul  9 11:11 ..
drwxrwxrwx  2 fog  fog  4,0K Jan 29 02:42 20160128W7PreSysprep
drwxrwxrwx  2 fog  fog  4,0K Jan 29 13:10 20160129W7PostSysprep
drwxrwxrwx  2 root root 4,0K Jul  9 19:37 dev
-rwxrwxrwx  1 root root    0 Jan 27 16:13 .mntcheck
drwxrwxrwx  2 root root 4,0K Mar 23 14:43 postdownloadscripts
drwxrwxrwx  2 fog  fog  4,0K Jul  7  2015 Windows7Software

$ ls -lha /mnt/olifant/images/
total 48K
drwxrwxrwx 12 root root 4,0K Jul  7 21:43 .
drwxr-xr-x  5 root root 4,0K Mar 23 18:01 ..
drwxrwxrwx  2 root root 4,0K Jan 27 16:16 20160127Temp
drwxrwxrwx  2 root root 4,0K Jul  7 10:45 20160706W7PreSysprep
drwxrwxrwx  2 root root 4,0K Jul  7 20:57 20160707W7PreSysprep
drwxrwxrwx  2 root root 4,0K Aug  1  2015 box20150731
drwxrwxrwx  2 root root 4,0K Jul  9 19:38 dev
drwxrwxrwx  2 root root 4,0K Apr  9  2015 lab2win720150409
drwxrwxrwx  2 root root 4,0K Apr 10  2015 lab2win720150410
drwxrwxrwx  2 root root 4,0K Jul 31  2015 lab420150716
-rwxrwxrwx  1 root root    0 Jan 27 16:13 .mntcheck
drwxrwxrwx  2 root root 4,0K Aug  1  2015 thin20150801
drwxrwxrwx  2 root root 4,0K Jan 26 15:45 Win7noDriversnoDefaultUser

[Wayne Workman]

@dolf said in Uploading to wrong storage group:

/images *(ro,sync,no_wdelay,no_subtree_check,insecure_locks,no_root_squash,insecure,fsid=0)
/images/dev *(rw,async,no_wdelay,no_subtree_check,no_root_squash,insecure,fsid=1)
/mnt/olifant/images *(ro,sync,no_wdelay,no_subtree_check,insecure_locks,no_root_squash,insecure,fsid=0)
/mnt/olifant/images/dev *(rw,async,no_wdelay,no_subtree_check,no_root_squash,insecure,fsid=1)

That’s the problem I think (or at least one of them).

fsid cannot have duplicates.

You have: 0,1,0,1
They should be: 0,1,2,3

[dolf]

Changed fsids as suggested, and rebooted. No change. To be honest, I have not worked with NFS before. Thanks for the help thus far.

[Wayne Workman]

@dolf Does tree list hidden files? The command I posted does. You should have these files in these directories:
/mnt/olifant/images/.mntcheck
/mnt/olifant/images/dev/.mntcheck
/images/.mntcheck
/images/dev/.mntcheck

[dolf]

A few posts ago, I did post the output of ls - lha which shows the existence of .hidden files. They are present with the right permissions, IMO, but I pasted the output in case I overlook something.

[Wayne Workman]

Also - temporarily disable the Koei (root) storage node. It’s a check box in it’s settings.

Then see what happens.

[dolf]

No change when disabling koei. Also disabled olifant, tried again, enabled one and tried again, then both and tried again. No change.

[Wayne Workman]

@dolf Why does the olifant node have the directory /mnt/... ? In it’s description, you say it’s in lab2-server, is that the fog server or another server? Have you mounted a remote directory to the fog server?

[dolf]

I have only one physical Ubuntu machine running fog, with two hard drives. There is a storage node on each, belonging to different storage groups, so that I don’t have to have all of the images on one disk (not enough space). /mnt/olifant is simply the mount point for a physical 1.5TB internal SATA drive with one ext4 partition.