Split handling HTTP requests into separate class
Until now HTTP requests were handled in the same class as incomingconnections (http.server.HttpServer). With this change the requesthandling is delegated to a separate class which can be re-used in tests...
Remove deprecated “QueryLocks” LUXI request
This has been deprecated since Ganeti 2.4 and hasn't been used byGaneti's code since. I'm not aware of any external users.
Signed-off-by: Michael Hanselmann <hansmi@google.com>Reviewed-by: Iustin Pop <iustin@google.com>
Add instance_balloon_memory rpc
Signed-off-by: Guido Trotter <ultrotter@google.com>Reviewed-by: René Nussbaumer <rn@google.com>
Accept both PUT and POST in noded
Currently, noded requires PUT, even though the semantics of the RPCcalls do not match a PUT. We change the code accept both PUT and POST,with the intention to remove the PUT support in a later version.
On the client side, the RPC code is changed to send POST requests....
rpc: Change signature of RpcRunner
Instead of receiving the whole masterd context, it only gets theconfiguration and the lock monitor callback. This simplifiesunittesting.
Additionaly, two new arguments for testing are added.
Signed-off-by: Michael Hanselmann <hansmi@google.com>...
Switch blockdev_getmirrorstatus_multi to per-node bodies
Even for a small 4-node node-group, this reduces the payload of thisRPC call by half for each node; for a big node group, previous testing(using a different patch) showed a general reduction in traffic over...
serializer: Remove JSON indentation and dict key sorting
Serializing to JSON using “simplejson” is significantly slower whenindentation and/or sorting of dictionary keys is used. In simplejson 1.xthe difference isn't that big, but with simplejson 2.x the difference...
Change “node_info” RPC to accept multiple VGs/hypervisors
Keeping the node state up to date will require information from multipleVGs and hypervisors. Instead of requiring multiple calls this changeallows a single call to return all needed information. Existing users...
masterd: Shutdown only once running jobs have been processed
Until now, if masterd received a fatal signal, it would start shuttingdown immediately. In the meantime it would hang while jobs are stillprocessed. Clients couldn't connect anymore to retrieve a jobs' status....
masterd: Initialize job queue only after RPC client
Otherwise jobs started after an unclean master shutdown will fail asthey depend on the RPC client.
masterd: Don't pass mainloop to server class
It is not used.
Change master IP address RPCs for external script
Change the master IP address RPC call chain to accept theuse_external_master_ip_script parameter. Introduces an unused parameterin backend.ActivateMasterIp and backend.DeactivateMasterIp, that will beused in the next commit....
Show RPC calls from config in lock monitor
With this patch all RPC calls at runtime of masterd will show up in thelock monitor. There is a chicken-and-egg issue with initializing theconfiguration with a context since the lock manager, containing themonitor, requires the configuration. This is worked around by setting...
Pass MasterNetworkParameters instances in RPCs
Pass instances of objects.MasterNetworkParameters when calling RPCs foractivation and deactivation of master IP.
Signed-off-by: Andrea Spadaccini <spadaccio@google.com>Reviewed-by: Guido Trotter <ultrotter@google.com>
Use MasterNetworkParameters attributes for RPC
Instead of manually unpacking the return values ofcfg.GetMasterNetworkParameters, let it return an instance ofobjects.MasterNetworkParameters and pass its attributes.
Signed-off-by: Andrea Spadaccini <spadaccio@google.com>...
Uniform master IP activation and deactivation
Add the master IP family parameter to the master IP deactivation RPCs,so that the activation and deactivation interfaces are uniform.
Explicitly pass params to activate_master_ip
To remove the usage of ssconf in the backend, the master needs to pushthe parameters of activate_master_ip to the backend.
This patch changes the entire call path of activate_master_ip to use thenew interface....
Change the semantics of stop_master RPC
To avoid polluting the stop_master RPC interface with networkparameters, this patch removes the deactivation of the master IP fromthis RPC.
The call to deactivate_master_ip is then added before every stop_master...
Explicitly pass params to deactivate_master_ip
Make the master explicitly pass the parameters to thedeactivate_master_ip RPC, and change all the call flow to use the newinterface.
Signed-off-by: Andrea Spadaccini <spadaccio@google.com>Reviewed-by: Michael Hanselmann <hansmi@google.com>
Explicitly pass params to change_master_netmask
Make the master explicitly pass the parameters to thechange_master_netmask RPC, and change all the call flow to use the newinterface.
Add new lock level for node resource modifications
This is in preparation for implementing a new resource model.
rpc: Change {import,export}_start to take source/dest in single argument
This simplifies the RPC argument encoding code.
Signed-off-by: Michael Hanselmann <hansmi@google.com>Reviewed-by: René Nussbaumer <rn@google.com>
rpc: Convert last two calls to generated code
These two calls, “upload_file” and “write_ssconf_files” are treatedseparated as they're used by the configuration, where we can't use thenormal resolver.
There's still some duplicated code in rpc.py, but that will be taken...
Move bootstrap-related RPC to generated wrappers
With this patch, only 6 RPC are left as old-style code.
Convert job queue's RPC to generated code
With these changes job queue RPC will finally show up on the lockmonitor. See below for an example. A job queue-specific class is used torestrict the use of a static list for name resolution to the job queue.Further improvements can be made to not re-create the whole RPC client...
rpc: Convert blockdev-related calls
Fix parameters to RPC "blockdev_rename"
Before this patch it would take a variable number of arguments, whichisn't very good for a definition-based approach.
Remove “node_tcp_ping” RPC
It is not used anywhere.
Standardise LUXI call argument types
Currently, we have 4 types of arguments in LUXI calls:
- most common, a list of values- a single argument that is sent as a list of one element- a single argument that is sent by itself- a dictionary (only Query and QueryFields)...
Rename filter and filter_ to qfilter
We currently use 'filter' as the OpCode, QueryRequest and RAPI fieldname for representing a query filter. However, since 'filter' is abuilt-in function, we actually have to use filter_ throughout the codein order to not override the built-in function....
Add cluster netmask parameter
Add the master_netmask cluster parameter, that represents the netmask ofthe master IP, encoded as a CIDR suffix.
This parameter can be set via the --master-netmask of gnt-cluster initand gnt-cluster modify. The default behaviour is to be consistent with...
Split starting and stopping master IP and daemons
Make migration RPC non-blocking
To add status reporting for the KVM migration, the instance_migrate RPCmust be non-blocking. Moreover, there must be a way to represent themigration status and a way to fetch it.
Keep only one global RPC runner in Ganeti context
Instead of having one RPC runner per mcpu processor this will keep onlyone instance as part of the masterd-wide Ganeti context. Upcomingpatches will change the RPC runner to report pending requests to the...
DeprecationWarning fixes for pylint
In version 0.21, pylint unified all the disable-* (and enable-*)directives to disable (resp. enable). This leads to a lot ofDeprecationWarning being emitted even if one uses the recommendedversion of pylint (0.21.1, as stated in devnotes.rst)....
PEP8 style fixes
Identified using the “pep8” utility.
Instance transfer: export component name to backend
This modifies the RPC layer to export the component name too to thebackend, so that it can be used in log files and messages.
Signed-off-by: Iustin Pop <iustin@google.com>Reviewed-by: Michael Hanselmann <hansmi@google.com>
Add gnt-instance start --pause
Creates the instance, but pauses execution before booting. This combinedwith 'gnt-instance console' unpausing instances means that the entireboot process can be viewed and monitored.
Signed-off-by: Stephen Shirley <diamond@google.com>...
Merge remote branch 'origin/devel-2.4'
Preload the string-escape code in noded
This encoding, part of the standard Python installation, is used bythe pickle module (in turn used by subprocess when handlingfailures in program execution). Preloading it means that Python willcache it in memory so that even if the disk goes away or just the...
Implement grow dry-run at RPC level
Signed-off-by: Iustin Pop <iustin@google.com>Reviewed-by: Guido Trotter <ultrotter@google.com>
Merge branch 'devel-2.4'
Fix master IP activation in failover with no-voting
Thanks to net.for.hub@gmail.com for reporting this. The logic inmasterd.CheckMasterd did an early return in case of no_voting, henceskipping the master IP activation. We just change the ifs to notreturn but simply continue through the function....
masterd: Simplify code for field queries
Instead of going via cmdlib and using special cases for differentresources, the list of fields is used directly.
constants: Rename QR_OP_*, add QR_VIA_RAPI
Commit 28b71a76 added a list of resources which can be queried usingLUXI. Unfortunately the variable was named “QR_OP_LUXI”, which can beconfusing. This patch renames “QR_OP_QUERY” to “QR_VIA_OP”, “QR_OP_LUXI”...
Add bdev_sizes RPC call
The bdev_sizes multi-node RPC call returns the sizes of the requestedblock devices on the desired nodes. Its intended use is to verify theexistence of a block device on a given node for shared block storagesupport.
Block device paths are expected to lie under constants.BLOCKDEV_DIR...
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...
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 OpQueryNodes and LUQueryNodes
Rename OpQueryInstances and LUQueryInstances
Signed-off-by: Iustin Pop <iustin@google.com>Reviewed-by: René Nussbaumer <rn@google.com>
Rename OpQueryGroups and LUQueryGroups
Rename OpQueryExports and LUQueryExports
Rename OpQueryConfigValues and LUQueryConfigValues
Rename OpQueryClusterInfo and LUQueryClusterInfo
Add RPC calls for pause/resume sync of devices
Signed-off-by: René Nussbaumer <rn@google.com>Reviewed-by: Iustin Pop <iustin@google.com>
Convert “gnt-debug locks” to query2
Locks can now be queried using “Query(what="lock", …)” over LUXI.
LUXI: Add Query and QueryFields functions
Ensure we just serialize JSON if there is data
locking: add nodegroup lock level
This also changes masterd to initialize the ganeti's manager with thecurrent list of nodegroup uuids, and updates unittests
Signed-off-by: Guido Trotter <ultrotter@google.com>Reviewed-by: Michael Hanselmann <hansmi@google.com>
Querying node groups: add luxi.REQ_QUERY_GROUPS
This also updates masterd.py.
Signed-off-by: Adeodato Simo <dato@google.com>Signed-off-by: Iustin Pop <iustin@google.com>Reviewed-by: Iustin Pop <iustin@google.com>
Adding RPC infrastructure to call oob
Fix disk checks in “gnt-cluster verify”
Tests have shown that the changes in commit b8d26c6e5 don't work aswanted. If any disk wasn't found on the node, all disks located on thesame node would show as faulty. The cause was incorrect exceptionhandling on the node....
Remove shebang line from ganeti.server.*
Some of then were forgotten.
Move ganeti-noded to ganeti.server.noded
Move ganeti-rapi to ganeti.server.rapi
Prepare move of daemons to ganeti.server
Move ganeti-masterd to ganeti.server.masterd
Move ganeti-confd to ganeti.server.confd