bdev: Add verification for file storage paths
An earlier version of this patch series verified all paths in cmdlib inthe master daemon. With this change all that verification code is movedto bdev to run inside the node daemon. The checks are much stricter...
jqueue: Factorize code to modify job
A new function will be added to change a job's priority.
Signed-off-by: Michael Hanselmann <hansmi@google.com>Reviewed-by: Bernardo Dal Seno <bdalseno@google.com>
cli: Use callback for --priority
If the option is used elsewhere, the numeric value is directlyavailable.
jqueue: Add docstring for _DetermineJobDirectories
Somehow this was missed in commit 0422250e.
Signed-off-by: Michael Hanselmann <hansmi@google.com>Reviewed-by: Helga Velroyen <helgav@google.com>
jqueue: Fix comments in _SubmitJobUnlocked
Drop SSHS_FORCE constant
It is not actually used.
Improve logging of AssertionErrors
Currently, when we have an assertion error raised from cmdlib, it looks like this:
[cluster] root@node4:~# gnt-instance grow-disk instance1 0 1G Failure: command execution error:
This is very very confusing. This patch adds a bit of traceback...
tools.prepare_node_join: Fix pep8 errors
Pep8 didn't agree with the indentation.
Signed-off-by: Michael Hanselmann <hansmi@google.com>Reviewed-by: Guido Trotter <ultrotter@google.com>
Add initial implementation of prepare-node-join
This is a new tool as per the design document “design-ssh-setup”. Itreceives a JSON data structure on its standard input and configures theSSH daemon and root's SSH keys accordingly. Unit tests are included....
ssh.GetUserFiles: RSA support, unit tests
This patch changes “ssh.GetUserFiles” to support two different kinds ofSSH keys, RSA and DSA. Before it would always use DSA. Newly writtenunit tests are included.
Signed-off-by: Michael Hanselmann <hansmi@google.com>...
Update blockdev's "info" at instance rename
Currently, we set "info" metadata on block devices at device creationtime, but we never update it, leading to stale data in case ofinstance renames. This would not be a big problem in case of regularrenames (assuming this is a rare operation), but importing instances...
LVM: remove old tags when adding new ones
This patch adds a small helper function to clear an LV's tags, andcalls it at SetInfo time. We need this to be able to correctly trackinstance renames, once we will call SetInfo at such times.
Signed-off-by: Iustin Pop <iustin@google.com>...
Add a small bdev helper function
I wanted to write that snippet the third time, which is too much :)
Signed-off-by: Iustin Pop <iustin@google.com>Reviewed-by: Michael Hanselmann <hansmi@google.com>
Compare significant fields only for simple SSH keys
For simple SSH keys, that is those without options such as“command="…"”, only the first two parts need to be compared. The thirdfield is a free-form comment.
This patch changes the comparison used in...
Merge branch 'devel-2.6'
Signed-off-by: Michael Hanselmann <hansmi@google.com>Reviewed-by: Iustin Pop <iustin@google.com>
ensure-dirs: Don't accept arguments
Before they would just be silently ignored.
ensure-dirs: Fix program name on usage screen
No string replacements are used, so doubling of the percent sign is notnecessary.
Before: Usage: %ensure-dirs [--full-run]After: Usage: ensure-dirs [--full-run]
Merge branch 'devel-2.6' into master
cli: Fix small typo
s/it/if/
Signed-off-by: Michael Hanselmann <hansmi@google.com>Reviewed-by: Dato Simó <dato@google.com>Reviewed-by: Helga Velroyen <helgav@google.com>
Group.hs: add 'allTags'; adjust loaders and test data for it
This commit adds a Group.allTags field to store the tags of node groups,and teaches each loader backend in HTools to populate it (additionally, theIAllocator class in lib/cmdlib.py now includes tags for groups too). Test...
Remove support for PUT in noded
This takes care of a FIXME; 2.6 already uses the new method, so we'regood during upgrades.
Ignore empty/comment lines in OS variants file
Per a conversation on ganeti@googlegroups.com:
- gnt-os diagnose ; gnt-os list take in consideration blank lines in /etc/ganeti/instance-image/variants.list that could be confusing.
Let's fix this and also let's ignore comment lines....
gnt-job cancel: Confirmation and selection of jobs
New parameters, “--pending”, “--queued” and “--waiting”, are added toselect all jobs in the respective state. If one of those options is usedand “--force” is not given, the user is asked to confirm the operation....
Add new constant for pending job status
This constant contains the job status' “queued”, “waiting” and“cancelled”.
Conflicts: NEWS: Trivial lib/tools/ensure_dirs.py: constant moved to pathutils...
ensure-dirs: Fix permissions on master socket
A socket shouldn't have its executable bit set.
errors: Document arguments to QueryFilterParseError
Also fix one small mistake in the docstring for QuitGanetiException.
Add support for cpu_cap and cpu_weight Xen params
This patch adds support for Xen's CPU scheduler 'cpu_cap' and'cpu_weight' parameters.
Ganeti default values (cap: 0=unlimited, weight: 256) are Xen defaults.
cpu_cap is not validated correctly because of actual Ganeti limitation...
LUClusterVerifyGroup: Localize virtual file paths
The check for file consistency didn't properly handle virtual pathsin case of a virtual cluster. This didn't cause any breakage as ina standard virtual cluster setup with only one node all files arevisible for every node....
Enable query socket usage in gnt-node/gnt-group
This switches gnt-node/gnt-group (and their equivalent RAPI resources)to go over the query socket.
vcluster: Don't virtualize /etc/hosts path
/etc/hosts is a bit special as it's a system-wide file and the virtualcluster/node root doesn't apply. The modification of /etc/hosts shouldbe disabled in virtual clusters. If it isn't, however, the vclusterfunctions would raise an exception complaining about a path outside of...
cli: Stop hardcoding /etc/hosts path
There is a constant for this purpose.
Move constant for /etc/hosts to pathutils
Needed for coming patches.
Merge branch 'stable-2.6' into devel-2.6
gnt-job list: Add option to include archived jobs
This provides a convenience option to include archived jobs in theoutput list. It's equivalent to using “-o +archived”, but tab completionis nicer.
gnt-job: List archived jobs if requested
If requested via a filter or by including the “archived” output,archived jobs will be loaded and shown. This is significantly slowerthan just listing normal jobs, therefore by default they are not loadedat all....
query: Report data type for unary operators
All data kinds (used to restrict the data collected) referenced in afilter can be requested once it's been “compiled”. However, the kindsof fields used in boolean expressions (e.g. ["?", "xyz"]) were notrecorded. This patch changes the code accordingly and provides a unit...
jqueue: Add new in-memory attribute for archived jobs
This attribute is set to True for jobs which were restored from anarchived file. A new filter will act on this field.
jqueue: Correct docstring
The description was not accurate.
verify-disks: Explicitely state nothing has to be done
Example output:$ gnt-cluster verify-disksSubmitted jobs 4327Waiting for job 4327 ...No disks need to be activated.
Add basic unit tests for "gnt-cluster epo"
This patch adds some unit tests for “gnt-cluster epo”. Not everything iscovered, but at least the bug fixed in the previous patch is.
Fix pylint breakage due to unused var in gnt_cluster
The usage of that variable was removed in 45a36f36, but accidentallythe enumerate() was left in.
cluster epo: Fix bug where IndexError is raised
Updating the “node_query_list” variable would fail if no arguments werepassed and the “--all” option wasn't specified. A follow-up patch willadd unit tests.
Fix usage of errors.ResolverError
This exception is documented to have three arguments, but in one casewe raise it with a simple string argument. Let's fix that.
Remove unused/deprecated error classes
It seems a few of the error classes are no longer used:
- LVMError, deprecated in 8c5533a5 (before ganeti 1.2.2!)- ConfdRequestError, deprecated in b0dcdc10- SshKeyError, introduced in the initial open source commit but never used (⁈)...
backend: Use utils.IsBelowDir instead of local code
utils.IsBelowDir is actually tested and doesn't allow writes to“…/queue*”, like the old code here did.
jstore: Nicer error message on non-numeric file content
An error like “invalid literal for int() with base 10” can be quiteconfusing.
bdev: Remove unnecessary empty line
My local pylint didn't complain.
Better list of replace-disks arguments + typos fixed
The man page and the bultin-in help for gnt-instance replace-disks wereinconsistent. Also fixed some typos in man pages.
Signed-off-by: Bernardo Dal Seno <bdalseno@google.com>Reviewed-by: Michael Hanselmann <hansmi@google.com>
Check fingerprint of file with allowed file storage paths
This makes differences show up in “gnt-cluster verify”.
Support for the default iallocator in replace-disks
"gnt-instance replace-disks" now behaves like the other commands, and usesthe default iallocator when "." is passed as the iallocator parameter.
Explicitly ask for the default iallocator in commands
Now "gnt-instance recreate-disks" uses the default iallocator when "." isspecified as the iallocator. For uniformity, the same behavior applies tothese commands: gnt-node evacuate gnt-instance migrate...
bdev: Add functions to verify file storage paths
- LoadAllowedFileStoragePaths: Loads a list of allowed file storage paths from a file- CheckFileStoragePath: Checks a path against the list of allowed paths
The unit test for “utils.IsBelowDir” is updated with cases which weren't...
jqueue: Look at archived jobs when watching
First: This enables the use of “gnt-job watch $id” for archived jobs.
Now, the reason for actually making this work is that duringsufficiently large group or node evacuations jobs are archived beforethe client gets to poll for their output. This led to situations where...
backend: Check for shared storage also
If normal file storage was disabled but shared storage enabled,“_TransformFileStorageDir” would still throw an exception.
in “opcodes._CheckStorageType” there's also a check, but I wasn't quitesure what the correct way of handling it was, so I added a TODO comment....
utils.FilterEmptyLinesAndComments: Return list
We don't use generators often and lists are easier to re-use.
Wipe added space when growing disks
This patch adds code to wipe newly added disk space when growing disksusing “gnt-instance grow-disk”. “New disk space” is defined as the deltabetween the old block device size (not necessarily equal to the amountrecorded in the configuration) and the new recorded size. Extra caution...
cmdlib._WipeDisks: Code formatting
- LogInfo takes *args, no need to replace values right away- Don't overwrite wipe_chunk_size right after it's been set
Factorize removing comments and empty lines from string
This will also be used for verifying the file storage directory.
ssconf: Fix mistake made in commit ee501db
Move a function from backend to ssconf
The “WriteSsconfFiles” function is used to write ssconf files. By movingit we can avoid importing backend into bootstrap. The latter is importedby CLI programs and backend doesn't have much to do with them.
Show old primary/secondary node on disk replacement
People unfamiliar with Ganeti's internals might be confused with thedifferent hostnames showing up later in the process.
cmdlib: Change wording of messages during disk wipe
Error messages don't need to say “please” and it's already obvious someinvestigation is needed. LogWarning already logs the message using“logging.error” internally.
Remove constant for disk wipe block size
It is dangerous to have this block size as a global constant as thatcould give the impression of it being easily changed. Doing so withoutfurther adjustments to how “dd” is called will lead to disks not beingwiped properly....
gnt-instance reinstall: Don't always exit with success
If one or more jobs failed the exit status should be set accordingly.
LUClusterVerify: Ignore /proc/drbd if DRBD is disabled
This fixes issue 190. The problem was that the check for DRBD wasenabled if LVM storage is used and didn't depend at all on whether DRBDis enabled.
cli: Exit with status 0 for --help
This patch is somewhat longer than was anticipated. Before, commands like“gnt-instance --help” would exit with the status code 1 (failure). Withthis patch, those commands exits with 0 (success) while unknown commandsstill print the usage screen and exit with status 1....
Remove the python confd server side code
In 2.7 we will only support the Haskell version, if enabled.
Since the original hconfd enabling was a bit hack-ish (copying overthe actual installed ganeti-confd, Python version), the Makefile.amchanges are a bit more involved than just the removal of the Python...
Remove option to select the Python version of confd
This will need some more configure.ac checks for required libraries,which will come later with the Haskell dependencies reorganisation.
Pylint cleanup
The previous commits introduced some pylint errors.
Signed-off-by: René Nussbaumer <rn@google.com>Reviewed-by: Iustin Pop <iustin@google.com>
Modify batch-create to use MultiAlloc
This is the CLI part of the OpInstanceMultiAlloc change. There is someroom for improvement especially with user handling, but it works for nowand the code got simplified significantly.
QA tests still needs to be written....
Make validation error message more useful
Provide a custom Validate for OpInstanceMultiAlloc
This is needed, as we're not strictly separating validation frompopulating the defaults. We do this all in one function. If we nowreceive a non complete OpInstanceCreate opcode in instances, we wouldrun into AttributeErrors on those....
ganeti-cleaner: Separate queue cleaning code
This code does not need to run as root, therefore it's better to splitit out. It is now run with the same permissions as the master daemon.
Stop hardcoding root user
Some parts of the code still use a hardcoded user name: root. This patchreplaces all with a constant specified at build time. The end goal is tomake it possible to run a Ganeti cluster without any special privileges(of course this will prevent some functionality from working)....
ensure-dirs: Don't hardcode ssconf file group
Otherwise chown(2) will fail when noded doesn't run as root.
InstanceRecreateDisks: new assertion for node locks
Make sure that all the nodes where a disk operation is peformed are locked.
Fix missing locks in InstanceRecreateDisks
In case of in-place recreate-disks operation, all node locks were releasedbefore starting the actual execution.
ensure-dirs: Don't convert list to tuple
Tuples are data structures, not containers.
Signed-off-by: Michael Hanselmann <hansmi@google.com>Reviewed-by: René Nussbaumer <rn@google.com>
Always_failover doesn't require --allow-failover anymore
If an administrator sets always_failover, it means that there is no needfor another explicit approval to failover instead of migrating.
Signed-off-by: Bernardo Dal Seno <bdalseno@google.com>Reviewed-by: Michael Hanselmann <hansmi@google.com>...
Adding RAPI client for instance multi allocation
It was very easy to refactor CreateInstance to repurpose some part of itfor the instance multi allocation request. So we do this.
Signed-off-by: René Nussbaumer <rn@google.com>Reviewed-by: Michael Hanselmann <hansmi@google.com>
Populate the instance dict with opcode infos
This is needed so we can load it as an opcode. However, this informationshould not be available on the client, so we inject that info when wereceive the request.
Provided unittest to verify behaviour.
Signed-off-by: René Nussbaumer <rn@google.com>...
Revert unintentional change of daemon log file names
Commit 3329f4dea6 unintentionally changed the filenames of all daemonlog files. This patch reverts part of those changes.
instance_info now returns vcpus
instance_info and all_instances_info calls had different per-instancereturn types, this commit fixes the mismatch.
Signed-off-by: Agata Murawska <agatamurawska@google.com>Reviewed-by: Iustin Pop <iustin@google.com>
Verify the node count just if the allocation was successful
Implement virtual cluster support in Python code
- pathutils: Prepend node-specific prefix path- RPC: Use virtual paths (see vcluster.py)- SSH: Pass environment variables, use destination's node directory when copying files using scp, use GANETI_HOSTNAME to determine hostname...
vcluster: Add more details to docstrings
As requested by René Nussbaumer. Hopefully things are more clear now.
Adding RAPI resource for multi-allocation
This is straightforward.
Adding multi-allocation logic
This patch adds the Logical Unit for the instance multi allocationopcode.
Adding the new opcode for multi-allocation
Skeleton for the multi-allocation opcode
Refactor shared code
We've some work which we need to do for the multi allocation request aswell, so lets refactor it to separate functions.
_ComputeNics is kinda ugly, as it reserves the MAC. I added a TODO tothink more about how we factor this behaviour out....
Fix REQ_RESULT for MultiInstanceAllocRequest
Fix a validation issue
We want to validate the result only if the iallocation run was asuccess.
Signed-off-by: René Nussbaumer <rn@google.com>Reviewed-by: Michael Hanselmann <hansmi@google.com>Reviewed-by: Bernardo Dal Seno <bdalseno@google.com>
netutils: Use virtual hostname if set
This enables the hostname seen by other code through Hostname.GetFqdn tobe changed through the environment variable “GANETI_HOSTNAME”.
Add new module for virtual clusters
This module will take care of managing paths for virtual clusters.Unittests are included (100% coverage).
bootstrap.SetupNodeDaemon: Stop hardcoding bind address
Commit b43dcc5a11 added support for IPv6. To have the node daemon bindto an IPv6 address on startup it changed the code to pass a static bindaddress (0.0.0.0 or ::0). If a user had configured another bind address...
hv_kvm: Use pu.CONF_DIR variable
… instead of using SYSCONFDIR directly.