utils.SetupLogging: Return function to reopen log file
This function can be used from a SIGHUP handler to reopen log files.Initial, simple unittests are included.
Signed-off-by: Michael Hanselmann <hansmi@google.com>Reviewed-by: René Nussbaumer <rn@google.com>
utils.SetupLogging: Make program a mandatory argument
It's passed in by most users (daemons, CLI scripts) and for the others(burnin, watcher) it certainly doesn't hurt, especially when usingsyslog.
Signed-off-by: Michael Hanselmann <hansmi@google.com>...
utils.log: Restrict I/O error handling coverage
The I/O error will occur while opening the file, not while addingand configuring the handler.
utils.log: Split formatter building into separate function
burner: Trivial code cleanup
- Use constant for exit value- Configure logging from main function, not from class' “__init__”
burnin: Reuse existing function for debug value
Instead of using its own, burnin can use cli.SetGenericOpcodeOpts.
Merge node groups from other cluster
Signed-off-by: Stephen Shirley <diamond@google.com>Reviewed-by: René Nussbaumer <rn@google.com>
Enforce that new node groups have unique names
Add _UnlockedLookupNodeGroup()
This allows calling of _UnlockedLookupNodeGroup() from withinAddNodeGroup()
Fix grammar of var naming
flatten is the verb, flattened is the adjective.
Signed-off-by: Stephen Shirley <diamond@google.com>Reviewed-by: Michael Hanselmann <hansmi@google.com>
Minor grammar fix in QuitGanetiException docstring
cluster-merge should refuse to merge own cluster
Also fix type of Merger.cluster_name from list to string. This wouldhave triggered an error in sshRunner if cluster keys were in use.
Signed-off-by: Stephen Shirley <diamond@google.com>Reviewed-by: Guido Trotter <ultrotter@google.com>
Introduce re-openable log record handler
This patch adds a new log handler class based on the standard library'sBaseRotatingHandler. This new class allows the log file to be re-opened,e.g. upon receiving a SIGHUP signal. The latter will be implemented in...
Re-create instance disk symlinks on activate
This patch implements recreation of instance disk symlinks when theactivate-disks operation is run. Until now, it was not possible tore-create these symlinks without stopping and starting or migrating aninstance as the RPC call where this is done was in instance startup...
Add RAPI resource for instance console
Signed-off-by: Michael Hanselmann <hansmi@google.com>Reviewed-by: Iustin Pop <iustin@google.com>
Export console information as query field
This makes it possible to get the console information via a LUXI query.
manpage: gnt-group remove cannot remove last group
Signed-off-by: Stephen Shirley <diamond@google.com>Signed-off-by: Iustin Pop <iustin@google.com>Reviewed-by: Iustin Pop <iustin@google.com>
ConfigWriter: simplify _UnlockedVerifyConfig
This just adds a 'cluster' local variable for reducing duplication.
Signed-off-by: Iustin Pop <iustin@google.com>Reviewed-by: René Nussbaumer <rn@google.com>
ConfigWriter: add checks for be/nd/nic params
This adds checking (in the configuration) for invalid be, nd and nicparams. The code is a bit tricky as nd params are at cluster,nodegroup and node level, nicparams are at cluster and nic level,whereas beparams are at cluster and instance level....
Add e1000 nic support for HVM
Closes issue: 130
Signed-off-by: Guido Trotter <ultrotter@google.com>Reviewed-by: Michael Hanselmann <hansmi@google.com>
Prevent removal of last node group
- Add check in ConfigWriter to prevent last node group from being removed- Tidy up error message a bit
Fix instance list for instances running multiple times
If for some reason (e.g. failed migration) one instance is runningon multiple nodes the output can become inconsistent. To get that errorand make it consistent between runs we make the call on the secondary...
Small QA fixes: groups via RAPI, cluster OOB
Add “cluster-oob” to sample configuration file. Don't run RAPI grouptests if disabled.
Merge branch 'devel-2.3' into devel-2.4
cluster verify: add hvparams verification
Currently, the validity of the hypervisor parameters is only checkedat init/modification time, and not in the cluster verify. This is bad,as it can lead to inconsistent state that is only detected when thenext modification (which can be unrelated) is made, leading to...
Remove dumb-allocator
- Remove the actual code- Remove mentions of it from iallocator.rst, and use hail instead- Also remove mentions of "etch-image" and use "debootstrap+default" - Mention htools as the reference implementation in iallocator.rst
Signed-off-by: Guido Trotter <ultrotter@google.com>...
Open other clusters' config in foreign mode
Add (unused) arg to _OfflineClusterMerge
cli._RunWhileClusterStoppedHelper.Call passes (self, *args) to functionscalled via cli.RunWhileClusterStoppedHelper(). The code in cluster-mergewas broken by commit d8aab233.
Signed-off-by: Stephen Shirley <diamond@google.com>...
Fix unittest breakage on Python 2.4/2.5
Commit 70b0d2a29 broke unittests on Python 2.4 and 2.5. Turns out thatPython 2.6 and above allow classes to be passed as custom test runners,whereas earlier versions don't.
Ensure all resources are used by RAPI client
Check for duplicate RAPI URIs and handlers
RAPI client: Wrap /2/redistribute-config resource
RAPI client: De-/activating instance disks
Add unittest for RAPI client's ModifyInstance
Watcher: Fix endless repair tries for broken secondary
In cases where secondary was offline and not evacuated watcher triedto activate-disks in an endless manner, but this is useless, as thesecondary is offline and therefore not responding to this approach....
Verify disks: increase parallelism and other fixes
The recent work on multi-VG support has converted LUClusterVerifyDisksinto doing serialised calls to each node, as each node can havedifferent VGs. This is suboptimal, especially for big clusters, where...
gnt-cluster verify-disks: fix VG name
Recent multi-VG work already exports the missing LV names as vg/lv,not simply lv. So the query and addition of the VG name in gnt-clusterverify-disks is redundant, and even wrong for non-default-VGinstances.
Signed-off-by: Iustin Pop <iustin@google.com>...
Deactivate disks: allow skipping hypervisor checks
In some cases (e.g. the hypervisor not running at all), we might wantto force disk deactivation, skipping the hypervisor checks. I believethis is not a good thing to do all the time, so this patch adds the...
Wait for master to become available on initialization
This is analogue to the existing check for a responsive node daemon.
Start all daemons on cluster initialization
At least ganeti-confd was not started. It got started a few minuteslater by ganeti-watcher. Also move one pylint disable to the effectiveline.
Clarify job processing order in admin guide
The fact that jobs don't necessarily execute in order has been sourcefor some confusion. Hopefully this update will clarify things.
Improve option descriptions
Also replace hardcoded “xenvg” with constant.
Remove two unused variables
Signed-off-by: Iustin Pop <iustin@google.com>Reviewed-by: Michael Hanselmann <hansmi@google.com>
Show hidden/blacklisted OSes in cluster info
Since we can blacklist/hide non-existing OSes (for preseeding), wecannot query easily the OSes themselves for this status. Hence weexport the entire lists in cluster info (which should be cheaper thangnt-os diagnose)....
Further man page updates for OS parameters
Also replace one UTF-8 char with the ASCII equivalent, not all Pandocversions support it.
Change the Makefile to use bash as SHELL
This is because we want, whenever we use sequences of commands, to setpipefail, otherwise detecting build failures is difficult.
Add documentation for OS parameters
Fix LUOSDiagnose and non-vm_capable nodes
This skips non-vm_capable nodes in the OS diagnose search, since suchOSes will not be used anyway on those nodes.
Rephrasing two error messages for auto promotion
Using auto_promote or auto-promote can lead to confusion on using theuser facing interfaces. While auto-promote is fine for CLI it's not forRAPI and vice-versa. This patch should eliminate this confusion....
lvmstrap: fix logic bug for partition reread
The if structure in CheckReread is broken, and makes partitions rereadbe full of race issues (esp. after updating them).
Also fix a small message bug.
storage: Check that mapper is either used or None
This is a followup patch to the one moving GetAllocatable out tomodule level.
Fix bug in “gnt-node list-storage”
LVM PV storage units would always show as allocatable, even when theyweren't. For some reason I have not been able to determine, the functionparsing the attributes (“_GetAllocatable”) was not even called and thelist opcode simply returned the attribute string as the value (e.g....
Fix payload check for out-of-band health
This logic error was not detected before as health has not beenimplemented on the cli and therefore no QA code existed for that.
Signed-off-by: René Nussbaumer <rn@google.com>Reviewed-by: Iustin Pop <iustin@google.com>
Fix premature abort of LUOobCommand due to result.Raise
This is a bug I recognized while doing tests on gnt-node health. A leftover result.Raise line causes premature abort of LUOobCommand on thefirst node failing the RPC call. This is not expected behaviour for...
ht: Add TMaybeDict check
This replaces a number of equal “ht.TOr(ht.TDict, ht.TNone)” checks.
QA: replace ping with fping
This allows a lot of simplification in the TestIcmpPing, as fping cantake multiple arguments so we don't need anymore to create manycommands joined with &&.
Modify LUOobCommand to support multiple nodes
This will change the result of this LU to a query like result. A list oftuples with information about the state of the data.
It also includes the modification to the commands calling this opcode.
Signed-off-by: René Nussbaumer <rn@google.com>...
Merge branch 'devel-2.4'
Rename QRFS_* to RS_*
This patch renames QRFS_* to RS_* fields so they can be used in otherplaces (i.e. LUs) without confusion, as this was initially meant forquery operations.
Another fix for LUClusterVerifyDisks
The LVM queries should only be done for vm_capable nodes. In order todo this, we also add a new ConfigWriter method to abstract that query.
QA: also run gnt-cluster verify-disks
The bug recently reported by Apollon Oikonomopoulos was missed becausewe don't test this command at all.
Fix disk adoption breakage
Disk adoption is currently broken by 84d7e26b, which added multiple LVMvolume group support. This patch fixes the calls to rpc.call_vg_list,which are multi-node calls but were handled as single-node calls in84d7e26b.
Signed-off-by: Apollon Oikonomopoulos <apollon@noc.grnet.gr>...
Fix typo in query2 design document
Improve import/export timeout settings
With this patch, the exporting node will retry to connect a few times.The receiving node will make use of the master's increased timeout (seeprevious patch).
Increase remote import/export timeout
It's been shown that 60 seconds may not be enough to establish aconnection.
lvmstrap: fix very old contact information
A memory from the past: this was left over from before the 1.2 releaseor so…
lvmstrap: add more excluded FS types
Also moves the list of excluded types to the top level and make it afrozenset.
lvmstrap: add an explicit test for mounted devices
Recent kernels/userland report a mounted filesystem as follows:
root@node4:~# fuser -avm /dev/sda5 USER PID ACCESS COMMAND /dev/sda5: root kernel mount /srv/ganeti...
lvmstrap: add explicit test for swap backends
Similar to mounted filesystems, recent kernel/userland report swapbackends:
root@node4:~# fuser -avm /dev/sda6 USER PID ACCESS COMMAND /dev/sda6: root kernel swap /dev/sda6...
lvmstrap: ignore small-sized partitions
This patch changes lvmstrap to ignore “small” partitions. Currentlyextended partitions are reported as unused as with a size of 1024(bytes), and this confuses lvmstrap. Since a very small partitionwon't help anyway (below hundred of PE size is not helpful), let's...
lvmstrap: abstract a little the sysfs paths
lvmstrap: add PV-on-partition support
This is a not-so-nice change, adding support for partitions to be usedas PVs.
The not-nice part is that partitions live in a separate place insysfs, whereas in dev they live at the same level as disks. Weworkaround this via a new SysfsName function that computes the correct...
Improve documentation for QRFS_UNAVAIL
IMHO this should have been named QRFS_NA or QRFS_UNSUPPORTED, butUNAVAIL is good enough.
query: Add alias support in _PrepareFieldList
Instance query: replace duplicates with aliases
Adding a basic oob helper as an example
This is just a plain stupid and simple out-of-band helper withoutanything fancy. It uses plain ssh to power off / power cycle themachine, does not support power on. It support power status using fpingto check if the host replies....
Fix disk count check in LUSetInstanceParams
LUSetInstanceParams checked instance.nics (and not instance.disks)against constants.MAX_DISKS.
Signed-off-by: Apollon Oikonomopoulos <apollon@noc.grnet.gr>Reviewed-by: Michael Hanselmann <hansmi@google.com>
Document iallocator change (alloc_policy)
Signed-off-by: Balazs Lecz <leczb@google.com>Reviewed-by: Michael Hanselmann <hansmi@google.com>
Add check-news to EXTRA_DIST
This was missing from commit 7385c51d.
Add script checking release dates in NEWS
This will detect human errors when setting a release date in NEWS.
query: use the actual types for BE/HV parameters
This patch exposes the VTYPE kind of BE/HV parameters, instead ofreturning QFT_OTHER. The current situation makes a query like:
gnt-instance list -o name,be/memory,oper_ram
very strange looking.
query: return UNAVAIL for "wrong" HV parameters
If a HV parameter is required that does not apply for an instance,currently the code returns None. This is bad, as it means we cannotswitch to the actual HV parameter types and validate correctly thisfield....
query: change (debug-mode) field validation errors
Currently, the single assert just checks that the entire row isconsistent (true/false), and dumps the row and field definitions as anaccompanying message. This makes it very hard to understand what failed....
Rename a few instance field titles
The conversion to querylib introduced unique field titles, which howeverdid an inconsistent rename of the instance vcpus/memory fields.
This patch makes both BE_MEMORY and VCPUS be named with prefix “Config”(which before was “Configured”), and drops the “Runtime” (note no...
query: Change internal result computation
While looking at the query library, I realized that while we have fivefield statuses, making this a 5-dimensional space, four of them areshrunk to a single possible value (None). Hence it should be possible toconvert this into a single value space plus extra 4 special constants....
Import QRFS_* and QFT_* in query.py
Currently, the QRFT/QFT constants are used with the module (e.g.constants.QRFS_NORMAL) in the query library. Given that there are 176uses of these constants in a module that has roughly ~1200 lines intotal (including whitespace), it makes sense to import these directly to...
Build epydoc's module list from Makefile
epydoc: specify files in tools/
This way we can add there non-python scripts
Signed-off-by: Guido Trotter <ultrotter@google.com>Reviewed-by: Iustin Pop <iustin@google.com>
KVM: Add auxiliary functions to handle tap devices
Add two new functions, _OpenTap and _ProbeTapVnetHdr, tohypervisors.hv_kvm.
_ProbeTapVnetHdr checks if the host kernel supports the virtio net header(IFF_VNET_HDR) feature. This feature allows KVM to ask the the host kernel to...
Ship default kvm-ifup script
In order to retain backwards compatibility, we ship the shell script previouslygenerated by _WriteNetScript as a stand-alone script under PKGLIBDIR.
Signed-off-by: Apollon Oikonomopoulos <apollon@noc.grnet.gr>Signed-off-by: Guido Trotter <ultrotter@google.com>...
KVM: Perform network configuration in Ganeti
This patch introduces network configuration for KVM in Ganeti.
There are three problems with having KVM perform network configuration via ifupscripts: a) Ganeti never gets to know the tap interface that is associated with an...
KVM: don't fail removing non-existing nic path
This shouldn't normally happen, but if it does no need to fail.
Rename OpGetTags and LUGetTags
Signed-off-by: Iustin Pop <iustin@google.com>Reviewed-by: René Nussbaumer <rn@google.com>Reviewed-by: Michael Hanselmann <hansmi@google.com>
Rename OpSearchTags and LUSearchTags
Rename OpAddTags and LUAddTags
Rename OpTestJobqueue and LUTestJobqueue
Check consistency of the class names and OP_ID
As the class names should be now consistent with the OP_IDs, we add acheck for wrongly-defined OP_IDs.
However, the future removal of the hand-coded OP_IDs will render thisobsolete, so this check is introduced just to make sure that the...
Remove hard-coded OP_ID definitions
The removal requires adding a few pylint ignores, as it cannot know thatOP_ID will be defined at runtime.
Also switch the asserts to check for non-declaration.
Rename OpDiagnoseOS and LUDiagnoseOS
Rename OpDelTags and LUDelTags