- 09 Nov, 2014 2 commits
-
-
Mike Hibler authored
We still use realpath to validate the path up front, but we pass the original (DB) path on to the client-side. Passing the resolved path was wrong anyway for clients that write images across NFS, because the path the client uses could be different than that computed on the server (e.g., /proj/foo vs. /.amd_mnt/ops/proj/foo) due to the way mounts are done. Note that the server will again validate the client-provided path, so if someone were to mess with a symlink in the path between when create_image verifies it and when it gets used, there is still no danger. This will probably eliminate the need for the AMD hack, but I'll leave it just to be safe.
-
Mike Hibler authored
-
- 08 Nov, 2014 1 commit
-
-
Mike Hibler authored
-
- 07 Nov, 2014 4 commits
-
-
Mike Hibler authored
If an available partition device (aka, the 4th partition on the system disk) represents less than 5% of the spare space we have found, ignore it. This will allow us to continue to use the 4th partition on the system disk of the d710s (450GB or so) and the second disk (250GB), but not use the 2nd partition (3GB), which would make us thrash about on the system disk even more than usual. Mostly this is for the new HP server boxes, so it doesn't pick up the 10GB left over on the (virtual) system disk when we have 21TB available on the second (virtual) disk. Another hack til blockstores rule the world...
-
Mike Hibler authored
-
Mike Hibler authored
-
Leigh B Stoller authored
-
- 06 Nov, 2014 2 commits
-
-
Leigh B Stoller authored
treated like normal hosts (tmcd already was changed). This is fine for XEN, but will break OpenVZ, but we can burn that bridge later.
-
Leigh B Stoller authored
-
- 05 Nov, 2014 12 commits
-
-
Leigh B Stoller authored
-
Leigh B Stoller authored
-
Kirk Webb authored
Also loads a couple of Infiniband modules so that Infiniband tools work properly.
-
Leigh B Stoller authored
-
Leigh B Stoller authored
-
Leigh B Stoller authored
-
Leigh B Stoller authored
-
Leigh B Stoller authored
-
Jonathon Duerig authored
-
Leigh B Stoller authored
-
Leigh B Stoller authored
-
Leigh B Stoller authored
-
- 04 Nov, 2014 9 commits
-
-
Leigh B Stoller authored
-
Leigh B Stoller authored
-
Leigh B Stoller authored
-
Leigh B Stoller authored
descriptor). Stash the current typelist into the descriptor being cloned, just before cloning (safety measure, until I fix the image update path).
-
Leigh B Stoller authored
existing image in emulab-ops. This is a last resort, as we do store a copy of the typelist in the descriptor for just this purpose, but stuff happens.
-
Leigh B Stoller authored
-
Leigh B Stoller authored
version number.
-
Leigh B Stoller authored
usage: runsonxen [-p <parent>] <imageid> usage: runsonxen -a [-p <parent>] usage: runsonxen -c <imageid> Options: -n - Impotent mode -c - Clear XEN parent settings completely -a - Operate on all current XEN capable images -p - Set default parent; currently XEN43-64-STD
-
Leigh B Stoller authored
Mark image update time when importing; more important to know when the image was brought in then when it was updated on the remote site.
-
- 03 Nov, 2014 7 commits
-
-
Brecht Vermeulen authored
e.G. the following assign violations gave just 'Not enough bandwdith' as the 0 always matched. Now we match with any number 1-9 BEST SCORE: 50.5 in 0 iters and 0.008631 seconds unassigned: 0 pnode_load: 26 no_connect: 0 link_users: 0 bandwidth: 0 desires: 0 vclass: 0 delay: 0 trivial mix: 0 subnodes: 0 max_types: 0 endpoints: 0
-
Leigh B Stoller authored
out the group changes. This is the wrong way to do this, but a real fix will wait until I have to deal with it in CloudLab.
-
Leigh B Stoller authored
much quicker.
-
Leigh B Stoller authored
-
Leigh B Stoller authored
-
Leigh B Stoller authored
-
Leigh B Stoller authored
-
- 01 Nov, 2014 1 commit
-
-
Mike Hibler authored
-
- 31 Oct, 2014 1 commit
-
-
Jonathon Duerig authored
-
- 30 Oct, 2014 1 commit
-
-
Leigh B Stoller authored
But remain backwards compat with existing certificates and credentials.
-