- Router VMs don't have a chain rule with -def suffix, this fixes name and
properly removes VR vms not running on a host
- Before trying to remove dnats, filter empty/None elements from list
- destroy_ebtables_rules should check what kind of action is request to be
performed (-A for add or -D for removed) and execute based on that
- Before executing any command, log it for debugging purposes
- Method to cleanup bridge, may be used in future
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
The SG python script depends on ebtables-save which is not available on Debian
based distros (Ubuntu and Debian for example). The commit uses /proc/modules
to find available bridge tables (one of nat, filter or broute) and then
find VMs that need to be removed. Further it uses set() to remove duplicate VMs
so we don't try to remove a VM's rules more than once leading to unwanted errors
in the log.
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
The only artifact resolved from libvirt.org was org.libvirt:libvirt:0.5.1
this artifact is now available from maven's default central repository
This closes#180
Signed-off-by: Laszlo Hornyak <laszlo.hornyak@gmail.com>
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
(cherry picked from commit 9cf31b07144302d053192265f9dfb731a2daefe9)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
This fixes the issue of Security Groups not working in case of XenServer 6.5;
- Uses nethash ipset data-structure to store CIDRs (efficient than iphash and
avoids overflow errors in case users add /8 /4 ingress/egress cidrs)
- Support for ipset versions both on 6.2 and 6.5, both have different outputs. This
fixes the issue of destroy_network_rules_for_vm failing
- Implements defensive filtering of list, instead of popping last item without
checking if it's None or empty
- Greps using names that are 'quoted' to avoid bash errors
- Before setting up new network rule, tries to clean and remove old ipset entry
- Idents, whitespace and naming fixes
PS. This is my 1000th commit to the 🐵 project :)
This closes#186
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
This is a defensive enhancement for KVM SG script that filters out empty string
instead of popping last item which may or may not be an empty string.
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Revert back to TRACE logging while retrieving and recycling VMware context from the context pool.
(cherry picked from commit 0f224c858fd3612c66632ba3476bfc3f1f71e1df)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
While attaching a new disk to an instance, the unit number on the controller key should be the lowest unit number
that is not in use. And in case the controller type is SCSI it shouln't be the reserved SCSI unit number.
(cherry picked from commit f420dd55fb5c4ec40aec0232618b2162c6d2f88c)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Set destination volume path as NULL while duplicating volume during migration.
If migration fails, destination volume will be marked as removed. And if migration succeeds, volume path will be rightly updated.
(cherry picked from commit d30d5644bb59898ac60884280bd0767bf429ede8)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
MigrateVMWithVolumes-
1. If ESXi host version is below 5.1, ensure destination datastore(s) is mounted on the source host, then migrate the storage and then finally migrate the VM.
If destination storage(s) is not mounted on the source host,
- In case of NFS storage mount the storage(s).
- In case of VMFS storage fail the request for migration.
2. If EXi host version is 5.1 or above, simultaneously migrate the VM and its storage to the destination host and storage(s) respectively for both NFS and VMFS storage.
(cherry picked from commit adc836cc5eadb0b05aefa6fcdcf8c03841282609)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Correctly register node info for a newly created VMware context.
(cherry picked from commit 13bdc1cef4ee0fae2f263777317c170302c49689)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
As recently discussed on the dev list:
This sorts the domain lists based on their path.
Especially handy when having a lot of domains,
like in a public cloud.
Cherry-picked to 4.5 branch since commit exists in 4.4 and master branch.
(cherry picked from commit befa28251d0e250f07c52a9a2cd4441cc8bd86d8)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
As suggested by Wido on the dev ML changing the repo to eu.ceph.com to avoid
build failures. Will revert if ceph.com is up again.
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
- pool allocation checks for both root and data disks
- NPE checks to not add null object in collection or try to migrate null VM
- HA work tries need to increment and be given up when max retries are crossed
- VM creation should check IP address format for IPv4 and IPv6
- If userdata is not supported by a network, then fail early if userdata, ssh key,
or password enabled template is passed/used
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
CentOS 7 does not ship with ifconfig anymore. We should use ip commands instead.
This also works on older versions, like CentOS 6 and Ubuntu 12.x/14.x, that we
support.
(cherry picked from commit ac06ec02eb442a4cc218b21e95f8d9a2a9ccb446)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Conflicts:
plugins/hypervisors/kvm/src/com/cloud/hypervisor/kvm/resource/IvsVifDriver.java
Upgrades from 4.4 where not working properly due to errors like this:
ERROR 1833 (HY000): Cannot change column 'id': used in a foreign key constraint 'fk_global_load_balancing_rules_region_id' of table 'cloud.global_load_balancing_rules'
The disabling of the checks is for the current session only, so it's safe.
In the 4.4 branch, this file used to be at:
client/target/utilities/scripts/db/db/schema-442to450.sql
It's now moved. We need to edit it in 4.4 and 4.5 to have it effective.
This closes#164
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
(cherry picked from commit 8f4abbc7bd65e7a7cad67ba64b29d745aa0869c0)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
In 6ac06e5e5e3ceed4a3e3a86ea5f82ffb59c266f2 logrotate was changed to run hourly.
Some logrotate configs still have set `daily` only which results in logs not
rotated hourly. The only way to ensure the log is rotated is to use size.
This closes#162
(cherry picked from commit 0ada08aa85b305ef931298654f7b41a7db28b03b)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
- use sticky chmod 1777 on the mountpoint
- remove dead code
- port improved code for moving disk into corresponding folder from master
- for dummy worker case, check case for powered off vm state
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
It causes extreme long vm start time when managment server has slow connection to secondary storage
(cherry picked from commit 49c01e2ae4eb164438f78f9686813497d7fd162f)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
While looking for an ongoing VM snapshot task, check the task status to identify if the task is still running.
(cherry picked from commit 25a4f0dc530951e87fe357313dc983cb75aa8972)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Handling spaces in datastore name while extracting vmdk base name
Signed-off-by: Sateesh Chodapuneedi <sateesh@apache.org>
(cherry picked from commit aa84b05491007567b81b33443fd9f99f6992e52c)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Update the path and size of data volumes after snapshot creation/deletion by correctly trimming only the snapshot postfix of a disk.
(cherry picked from commit ac491c96075d65e35157380fe7d28fdd917c0e90)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Fix the OVA path that is returned once an OVA is packaged using a META file.
(cherry picked from commit 507d9d337d2e80f7f6e5d665708c0687c387c5fc)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
