Manual:CHR: Difference between revisions

From MikroTik Wiki
Jump to navigation Jump to search
← Older edit
No edit summary
 
(14 intermediate revisions by 6 users not shown)
Line 1: Line 1:
{{Warning|This manual is moved to https://help.mikrotik.com/docs/x/mgAYAQ}}
== Cloud Hosted Router ==
== Cloud Hosted Router ==


Line 6: Line 9:


====  Minimal requirements: ====  
====  Minimal requirements: ====  
* RouterOS version 6.34 or later installed
* Package version: RouterOS v6.34 or newer
* 64bit CPU with virtualization support
* Host CPU: 64-bit with virtualization support
* 128 MB or more RAM for the CHR instance
* RAM: 128MB or more
* 128 MB disk space for the CHR virtual hard drive  
* Disk: 128MB disk space for the CHR virtual hard drive (Max: 16GB)
** Maximum supported system disk image size is 16GB


{{Note| Minimal requirement is 128MB of RAM to complete the self-installation process.}}
{{Note| The minimum required RAM depends on interface count and CPU count. You can get an approximate number by using the following formula: RAM = 128 + [ 8 × (''CPU_COUNT'') × (''INTERFACE_COUNT'' - 1) ]}}


==== CHR has been tested on the following platforms: ====  
==== CHR has been tested on the following platforms: ====  
Line 76: Line 78:
* [[Manual:CHR_Hyper-V_installation|Hyper-V]]
* [[Manual:CHR_Hyper-V_installation|Hyper-V]]
* [[Manual:CHR_AWS_installation|Amazon Web Services (AWS)]]
* [[Manual:CHR_AWS_installation|Amazon Web Services (AWS)]]
* [[Manual:CHR_Hetzner|Hetzner Cloud]]
* [[Manual:CHR_Linode|Linode]]
* [[Manual:CHR_Linode|Linode]]
* [[Manual:CHR_GCE|Google Compute Engine]]
* [[Manual:CHR_ProxMox_installation|ProxMox]]
* [[Manual:CHR_ProxMox_installation|ProxMox]]


Line 133: Line 137:
To request a trial license, you must run the command "'''/system license renew'''" from the CHR device command line. You will be asked for the username and password of your mikrotik.com account.
To request a trial license, you must run the command "'''/system license renew'''" from the CHR device command line. You will be asked for the username and password of your mikrotik.com account.


{{Warning|If you plan to use multiple virtual systems of the same kind, it may be possible that the next machine has the same systemID as the original one. This can happen on certain cloud providers, such as Linode. To avoid this, after your first boot, run the command "/system license generate-new-id" '''before''' you request a trial license. }}
{{Warning|If you plan to use multiple virtual systems of the same kind, it may be possible that the next machine has the same systemID as the original one. This can happen on certain cloud providers, such as Linode. To avoid this, after your first boot, run the command "/system license generate-new-id" '''before''' you request a trial license. Note that this feature must be used only while CHR is running on free type of RouterOS license. If you have already obtained paid or trial license, do not use regenerate feature since you will not be able to update your current key any more}}


== Getting the License ==
== Getting the License ==
Line 190: Line 194:
==== Upgrade from higher tier up ====
==== Upgrade from higher tier up ====


Only an upgrade to a higher tier is possible at the moment (for paid licenses only) and that is done in the account server. For changes to take place on the router itself ''renew'' command should be issued. When the router already has any kind of trial or paid license, the license level you set for the ''renew'' command is not important anymore, it is mandated by [https://www.mikrotik.com/client/cloudhosting the account server]. Possible upgrades are as follows:
Only an upgrade to a higher tier is possible at the moment (for paid licenses only) and that is done in the account server. For changes to take place on the router itself ''renew'' command should be issued. When the router already has any kind of trial or paid license, the license level you set for the ''renew'' command is not important anymore, it is mandated by ['''license.mikrotik.com''' account server]. Possible upgrades are as follows:
* ''p1'' upgrade to ''p10''
* ''p1'' upgrade to ''p10''
* ''p1'' upgrade to ''p-unlimited''
* ''p1'' upgrade to ''p-unlimited''
Line 491: Line 495:


* [https://pve.proxmox.com/wiki/Qemu-guest-agent Enabling guest agent on proxmox]
* [https://pve.proxmox.com/wiki/Qemu-guest-agent Enabling guest agent on proxmox]
* Enabling 'chr.provision-agent' for remote access on port 1234
In host shell:
<pre>
vmid=256
hostip=192.168.0.1
portnum=1234
qm set $vmid --args "-chardev 'socket,host=$hostip,port=$portnum,id=chr-agent,server,nowait' -device 'virtio-serial,bus=pci.0,addr=0x9' -device 'virtserialport,chardev=chr-agent,name=chr.provision_agent'"
</pre>
* Disabling 'chr.provision-agent'
In host shell:
<pre>
vmid=256
qm set $vmid --delete args
</pre>
* Providing remote access to default agent on port 1234
In host shell:
<pre>
vmid=256
portnum=1234
socat TCP-LISTEN:$portnum,reuseaddr,fork UNIX-CLIENT:/run/qemu-server/$vmid.qga
</pre>
==== Guest agent python example ====
<pre>
import os,time,base64,json,socket,select,errno
class GuestAgent(object):
    '''
        Qemu guest agent interface
        runScript and runFile commands are tailored for ROS agent implementation
        Transport provided by derived classes (transact method)
    '''
    def __init__(self,**kwargs):
        # Due to file contents being passed as base64 inside json:
        #  - large chunk sizes may slow down guest-side parsing.
        #  - small chunk sizes result in additional message fragmentation overhead.
        # Default value is a guestimate.
        self.__chunkSize = kwargs.get('chunkSize', 4096)
    def _qmpError(self,cls,msg):
        ''' Generic callback to log qmp errors before (optionally) raising an exception '''
        print(cls)
        for line in msg.split('\n'):
            print(line)
        # raise RuntimeError()
    def _error(self,msg,*a):
        ''' Generic callback to misc errors before (optionally) raising an exception '''
        print(msg.format(*a))
        # raise RuntimeError()
    def _info(self,msg,*a):
        ''' Generic callback to log info '''
        print(msg.format(*a))
    def _monitorJob(self,pid):
        ''' Block untill script job completes, echo output. Returns None on failure '''
        ret = self.transact('guest-exec-status',{'pid':pid})
        if ret is None:
            return None
        while not bool(ret['exited']):
            time.sleep(1)
            ret = self.transact('guest-exec-status',{'pid':pid})
            if ret is None:
                return None
        # err-data is never sent
        out = []
        if 'out-data' in ret.keys():
            out = base64.b64decode(ret['out-data']).decode('utf-8').split('\n')
            if not out[-1]:
                out = out[:-1]
        exitcode = int(ret['exitcode'])
        return exitcode, out
    def putFile(self,src,dst):
        ''' Upload file '''
        src = os.path.expanduser(src)
        if not os.path.exists(src) or not os.path.isfile(src):
            self._error('File does not exist: \'{}\'', src)
            return None
        ret = self.transact('guest-file-open', {'path':dst,'mode':'w'})
        if ret is None:
            return None
        handle = int(ret)
        file = open(src, 'rb')
        for chunk in iter(lambda : file.read(self.__chunkSize), b''):
            count = len(chunk)
            chunk = base64.b64encode(chunk).decode('ascii')
            ret = self.transact('guest-file-write',{'handle':handle,'buf-b64':chunk,'count':count})
            if ret is None:
                return None
        self.transact('guest-file-flush',{'handle':handle})
        ret = self.transact('guest-file-close',{'handle':handle})
        return True
    def getFile(self,src,dst):
        ''' Download file '''
        dst = os.path.expanduser(dst)
        ret = self.transact('guest-file-open',{'path':src,'mode':'rb'})
        if ret is None:
            return None
        handle = int(ret)
        data = ''
        size = 0
        while True:
            ret = self.transact('guest-file-read',{'handle':handle,'count':self.__chunkSize})
            if ret is None:
                return None
            data += ret['buf-b64']
            size += int(ret['count'])
            if bool(ret['eof']):
                break
        ret = self.transact('guest-file-close',{'handle':handle})
        data = base64.b64decode(data.encode('ascii'))
        with open(dst,'wb') as f:
            f.write(data)
        return True
    def runFile(self,fileName):
        ''' Execute file (on guest) as script '''
        ret = self.transact('guest-exec',{'path':fileName, 'capture-output':True})
        if ret is None:
            return None
        pid = ret['pid']
        return self._monitorJob(pid)
    def runSource(self,cmd):
        ''' Execute script '''
        if isinstance(cmd,list):
            cmd = '\n'.join(cmd)
        cmd += '\n'
        cmd = base64.b64encode(cmd.encode('utf-8')).decode('ascii')
        ret = self.transact('guest-exec',{'input-data':cmd, 'capture-output':True})
        if ret is None:
            return None
        pid = ret['pid']
        return self._monitorJob(pid)
    def shutdown(self,mode='powerdown'):
        '''
            Execut shutdown command
            mode == 'reboot' - reboot guest
            mode == 'shutdown' or mode == 'halt' - shutdown guest
        '''
        ret = self.transact('guest-shutdown',{'mode':mode})
        return ret
class SocketAgent(GuestAgent):
    '''
        GuestAgent using unix/tcp sockets for communication.
    '''
    def __init__(self):
        GuestAgent.__init__(self,chunkSize= 32 * 65536)
    @staticmethod
    def unix(dev):
        ''' Connect using unix socket '''
        self = SocketAgent()
        self.__af = socket.AF_UNIX
        self.__args = dev
        self.__wait = False
        return self
    @staticmethod
    def tcp(ip,port,wait = True):
        ''' Connect using tcp socket '''
        self = SocketAgent()
        self.__af = socket.AF_INET
        self.__args = (ip,port)
        self.__wait = wait
        return self
    def __enter__(self):
        self._sock = socket.socket(self.__af, socket.SOCK_STREAM)
        if self.__wait:
            self._info('Waiting for guest ...')
            # Wait for hyper to create channel
            while True:
                try:
                    self._sock.connect(self.__args)
                    break
                except socket.error as e:
                    if e.errno == errno.EHOSTUNREACH or e.errno == errno.ECONNREFUSED:
                        time.sleep(1)
                    else:
                        self._sock.close()
                        raise
            #Wait for guest agent to initialize and sync
            while True:
                import random
                key = random.randint(0, 0xffffffff)
                msg = json.dumps({'execute':'guest-sync-delimited','arguments':{'id':key}},separators=(',',':'),sort_keys=True)
                self._sock.send(msg.encode('ascii'))
                self._sock.setblocking(0)
                response = b''
                if (select.select([self._sock],[],[])[0]):
                    response += self._sock.recv(65536)
                else:
                    raise RuntimeError()
                self._sock.setblocking(1)
                sentinel = b'\xff'
                response = response.split(sentinel)[-1]
                if not response:
                    time.sleep(3)
                    continue
                response = json.loads(response.decode('utf-8').strip())
                if 'return' in response.keys():
                    if int(response['return']) == key:
                        break
                time.sleep(3)
        else:
            self._sock.connect(self.__args)
        return self
    def __exit__(self,*a):
        self._sock.close()
    def transact(self,cmd,args={}):
        ''' Exchange a single command with guest agent '''
        timeout = 2
        msg = json.dumps({'execute':cmd,'arguments':args},separators=(',',':'),sort_keys=True)
        self._sock.send(msg.encode('ascii'))
        self._sock.setblocking(0)
        response = b''
        if (select.select([self._sock],[],[],timeout)[0]):
            response += self._sock.recv(65536)
        self._sock.setblocking(1)
        if not response:
            response = None
        else:
            if response[0] == 255: # sync
                response = response[1:]
            response = json.loads(response.decode('utf-8').strip())
            if 'error' in response.keys():
                self._qmpError(response['error']['class'],response['error']['desc'])
                response = None
            elif 'return' in response:
                response = response['return']
        return response
#-------------------------------------------------------------------------------
if __name__ == '__main__':
    script = [':log info "hello world";']
    ip = '192.168.0.1'
    port = 1234
    # can also use unix sockets
    #with SocketAgent.unix('/dev/something') as agent:
    with SocketAgent.tcp(ip, port) as agent:
        ret,out = agent.runSource(script)
        print('ret = {}'.format(ret))
        for line in out:
            print(line)
</pre>




Latest revision as of 10:13, 7 December 2020

Warning: This manual is moved to https://help.mikrotik.com/docs/x/mgAYAQ



Cloud Hosted Router

Cloud Hosted Router (CHR) is a RouterOS version intended for running as a virtual machine. It supports the x86 64-bit architecture and can be used on most of the popular hypervisors such as VMWare, Hyper-V, VirtualBox, KVM and others. CHR has full RouterOS features enabled by default but has a different licensing model than other RouterOS versions.

System Requirements

Minimal requirements:

  • Package version: RouterOS v6.34 or newer
  • Host CPU: 64-bit with virtualization support
  • RAM: 128MB or more
  • Disk: 128MB disk space for the CHR virtual hard drive (Max: 16GB)

Note: The minimum required RAM depends on interface count and CPU count. You can get an approximate number by using the following formula: RAM = 128 + [ 8 × (CPU_COUNT) × (INTERFACE_COUNT - 1) ]


CHR has been tested on the following platforms:

  • VirtualBox 5 on Linux and OS X
  • VMWare Fusion 7 and 8 on OS X
  • VMWare ESXi 6.5
  • Qemu 2.4.0.1 on Linux and OS X
  • Hyper-V on Windows Server 2008r2, 2012 and Windows 10 (Only Generation 1 Hyper-V virtual machine is supported at the moment)
  • Xen Project 4.6.5
  • Xen Server 7.1

Warning: Hypervisors that provide paravirtualization are not supported.


Usable Network and Disk interfaces on various hypervisors:

  • ESX:
    • Network: vmxnet3, E1000
    • Disk: IDE, VMware paravirtual SCSI, LSI Logic SAS, LSI Logic Parallel
  • Hyper-V:
    • Network: Network adapter, Legacy Network adapter
    • Disk: IDE, SCSI
  • Qemu/KVM:
    • Network: Virtio, E1000, vmxnet3 (optional)
    • Disk: IDE, Sata, Virtio
  • Xen Project:
    • Network: E1000, rtl8193, netfront
    • Disk: IDE, Sata
  • VirtualBox
    • Network: E1000, rtl8193
    • Disk: IDE, Sata, SCSI, SAS

Note: SCSI controller Hyper-V and ESX is usable just for secondary disks, system image must be used with IDE controller!


Warning: We do not recommend using E1000 network interface if better synthetic interface options are available on specific Hypervisor!


How to Install CHR

We provide 4 different virtual disk images to choose from. Note that they are only disk images, and you can't simply run them.

  • RAW disk image (.img file)
  • VMWare disk image (.vmdk file)
  • Hyper-V disk image (.vhdx file)
  • VirtualBox disk image (.vdi file)

Steps to install CHR

  1. Download virtual disk image for your hypervisor
  2. Create a guest virtual machine
  3. Use previously downloaded image file as a virtual disk drive
  4. Start the guest CHR virtual machine
  5. Log in to your new CHR. Default user is 'admin', without password

Please note that running CHR systems can be cloned and copied, but the copy will be aware of the previous trial period, so you cannot extend your trial time by making a copy of your CHR. However, you are allowed to license both systems individually. To make a new trial system, you need to make a fresh installation and reconfigure RouterOS.

Installing CHR

CHR Licensing

The CHR has 4 license levels:

  • free
  • p1 perpetual-1 ($45)
  • p10 perpetual-10 ($95)
  • p-unlimited perpetual-unlimited ($250)

60-day free trial license is available for all paid license levels. To get the free trial license, you have to have an account on MikroTik.com as all license management is done there.

Perpetual is a lifetime license (buy once, use forever). It is possible to transfer a perpetual license to another CHR instance. A running CHR instance will indicate the time when it has to access the account server to renew it's license. If the CHR instance will not be able to renew the license it will behave as if the trial period has ran out and will not allow an upgrade of RouterOS to a newer version.

After licensing a running trial system, you must manually run the /system license renew function from the CHR to make it active. Otherwise the system will not know you have licensed it in your account. If you do not do this before the system deadline time, the trial will end and you will have to do a complete fresh CHR installation, request a new trial and then license it with the license you had obtained.

LicenseSpeed limitPrice
Free1MbitFREE
P11Gbit$45
P1010Gbit$95
P-UnlimitedUnlimited$250

p1

p1 (perpetual-1) license level allows CHR to run indefinitely. It is limited to 1Gbps upload per interface. All the rest of the features provided by CHR are available without restrictions. It is possible to upgrade p1 to p10 or p-unlimited After the upgrade is purchased the former license will become available for later use on your account.

p10

p10 (perpetual-10) license level allows CHR to run indefinitely. It is limited to 10Gbps upload per interface. All the rest of the features provided by CHR are available without restrictions. It is possible to upgrade p10 to p-unlimited After the upgrade is purchased the former license will become available for later use on your account.

p-unlimited

The p-unlimited (perpetual-unlimited) license level allows CHR to run indefinitely. It is the highest tier license and it has no enforced limitations.

Free licenses

There are several options to use and try CHR free of charge.

free

The free license level allows CHR to run indefinitely. It is limited to 1Mbps upload per interface. All the rest of the features provided by CHR are available without restrictions. To use this, all you have to do is download disk image file from our download page and create a virtual guest.

60-day trial

In addition to the limited Free installation, you can also test the increased speed of P1/P10/PU licenses with a 60 trial.

You will have to have an account registered on MikroTik.com. Then you can request the desired license level for trial from your router that will assign your router ID to your account and enable a purchase of the license from your account. All the paid license equivalents are available for trial. A trial period is 60 days from the day of acquisition, after this time passes, your license menu will start to show "Limited upgrades", which means that RouterOS can no longer be upgraded.

If you plan to purchase the selected license, you must do it within 60 days of the trial end date. If your trial ends, and there are no purchases within 2 months after it ended, the device will no longer appear in your MikroTik account. You will have to make a new CHR installation to make a purchase within the required time frame.

To request a trial license, you must run the command "/system license renew" from the CHR device command line. You will be asked for the username and password of your mikrotik.com account.

Warning: If you plan to use multiple virtual systems of the same kind, it may be possible that the next machine has the same systemID as the original one. This can happen on certain cloud providers, such as Linode. To avoid this, after your first boot, run the command "/system license generate-new-id" before you request a trial license. Note that this feature must be used only while CHR is running on free type of RouterOS license. If you have already obtained paid or trial license, do not use regenerate feature since you will not be able to update your current key any more


Getting the License

After the initial setup a CHR instance will have a free license assigned. From there, it is possible to upgrade the license to a higher tier. Once you have a trial license all the work with the license is done on the account server where it is possible to upgrade license to a higher tier unless it is p-unlimited already.

Upgrade from free to p1 or higher

Initial upgrade from the free tier to anything higher than that incurs CHR instance registration on the account server. To do that you have to enter your MikroTik.com username and password and a desired license level you want to acquire. As a result, a CHR ID number will be assigned to your account on the account server and 60-day trial created for that ID. There are 2 ways to obtain a license - using WinBox or RouterOS command line interface:

Using WinBox (Sytem -> License menu):

CHR Licence Renew

CHR Licence Renew

Using command line interface: 

[admin@MikroTik] > /system license print 
  system-id: 6lR1ZP/utuJ
      level: free

[admin@MikroTik] > /system license renew 
account: mymikrotikcomaccount
password: *********************
level: p1 
  status: done
  
[admin@MikroTik] > /system license print 
        system-id: 6lR1ZP/utuJ
            level: p1
  next-renewal-at: jan/10/2016 21:59:59
      deadline-at: feb/09/2016 21:59:59

To acquire a higher level trial, set up a new CHR instance, renew the license and select the desired level.

To upgrade from a Trial license to Paid go to MikroTik.com account server and choose 'all keys' in Cloud Hosted Router (CHR) section:

Account server

You will be presented with a list of your CHR machines and licenses:

Account server

To upgrade from a Trial to a Paid license click 'Upgrade', choose the desired license level (it can be different than the level of the trial license) and click 'Upgrade key':

Account server

Choose the payment method:

Account server

It is possible to pay using account balance (deposit), credit card (CC), PayPal or using Balance (prepaid) key (if you have any).

Upgrade from higher tier up

Only an upgrade to a higher tier is possible at the moment (for paid licenses only) and that is done in the account server. For changes to take place on the router itself renew command should be issued. When the router already has any kind of trial or paid license, the license level you set for the renew command is not important anymore, it is mandated by [license.mikrotik.com account server]. Possible upgrades are as follows:

  • p1 upgrade to p10
  • p1 upgrade to p-unlimited
  • p10 upgrade to p-unlimited

License Update

In '/system license' menu router will indicate the time next-renewal-at when it will attempt to contact server located on licence.mikrotik.com. Communication attempts will be performed once an hour after the date on next-renewal-at and will not cease until the server responds with an error. If deadline-at date is reached without successfully contacting the account server, the router will consider that license has expired and will disallow further software updates. However, router will continue to work with the same license tier as before.

Troubleshooting

Running on VMware ESXi

Changing MTU

VMware ESXi supports MTU of up to 9000 bytes. To get the benefit of that, you have to adjust your ESXi installation to allow a higher MTU. Virtual Ethernet interface added after the MTU change will be properly allowed by the ESXi server to pass jumbo frames. Interfaces added prior to MTU change on the ESXi server will be barred by the ESXi server (it will still report old MTU as maximum possible size). If you have this, you have to re-add interfaces to the virtual guests.

Example. There are 2 interfaces added to the ESXi guest, auto-detected MTU on the interfaces show MTU size as it was at the time when the interface was added: 

[admin@chr-vm] > interface ethernet print 
Flags: X - disabled, R - running, S - slave 
 #    NAME           MTU MAC-ADDRESS       ARP       
 0 R  ether1        9000 00:0C:29:35:37:5C enabled   
 1 R  ether2        1500 00:0C:29:35:37:66 enabled

Using bridge on Linux

If Linux bridge supports IGMP snooping, and there are problems with IPv6 traffic it is required to disable that feature as it interacts with MLD packets (multicast) and is not passing them through. 

echo -n 0 > /sys/class/net/vmbr0/bridge/multicast_snooping

Packets not passing from guests

The problem: after configuring a software interface (VLAN, EoIP, bridge, etc.) on the guest CHR it stops passing data to the outside world beyond the router.

The solution: check your VMS (Virtualization Management System) security settings, if other MAC addresses allowed to pass if packets with VLAN tags allowed to pass through. Adjust the security settings according to your needs like allowing MAC spoofing or certain MAC address range. For VLAN interfaces, it is usually possible to define allowed VLAN tags or VLAN tag range.

Using vlans on CHR in various Hypervisors

In some of hypervisors before Vlans can be used on VMs they need to first be configured on hypervisor it self.

ESXI

Enable Promiscuous mode in port group or virtual switch that you will use for specific VM.

ESX documentation:

Hyper-V

Hyper-V documentation:

bhyve hypervisor

It wont be possible to run CHR on this hypervisor. CHR cannot be run as paravirtualized platform.

Linode

When creating multiple Linodes with the same disk size, new Linodes will have the same systemID. This will cause issues to get a Trial/Paid license. To avoid this, run the command /system license generate-new-id after the first boot and before you request a trial or paid license. This will make sure the ID is unique.


Some useful articles:

Specific vlan is untagged by nic interface:

Allow passing other vlans:

Guest tools

VMWare

Time synchronization

Must be enabled from GUI ('Synchronize guest time with host'). Backwards synchronization is disabled by default - if guest is ahead of host by more than ~5 seconds, synchronization is not performed

Power operations

  • poweron and resume scripts are executed (if present and enabled) after poweron and resume operations respectively.
  • poweroff and suspend scripts are executed before poweroff and suspend operations respectively.
  • If scripts take longer than 30 seconds or contain errors, the operation fails
  • In case of failure, retrying the same operation will ignore any errors and complete successfully
  • Failed script output is saved to file (e. g. 'poweroff-script.log', 'resume-script.log' etc)
  • Scripts can be enabled/disabled from hypervisor GUI ('run VMware Tools Scripts') or by enbaling/disabling scripts from console

Quiescing/backup

Guest filesystem quiescing is performed only if requested.

  • freeze script is executed before freezing the filesystem
  • freeze-fail script is executed if hypervisor failed to prepare for snapshot or if freeze script failed
  • thaw script is executed after snapshot has been taken
  • Script run time is limited to 60 seconds
  • freeze script timeouts and errors result in backup operation being aborted
  • FAT32 disks are not quiesced
  • Failed script output is saved to file (e. g. 'freeze-script.log', 'freeze-fail-script.log', 'thaw-script.log')

Guest info

Networking, disk, and OS info is reported to hypervisor every 30 seconds (GuestStats (memory) are disabled by default, can be enabled by setting 'guestinfo.disable-perfmon = "FALSE"' in VM config).

  • The order, in which network interfaces are reported, can be controlled by setting 'guestinfo.exclude-nics', 'guestinfo.primary-nics' and 'guestinfo.low-priority-nics' options. Standard wildcard patterns can be used.

Provisioning

Can use the ProcessManager from vim API to execute scripts. Python bindings are available

  • Main data structure: GuestProgramSpec
    • The workingDirectory and envVariables members are ignored
    • programPath must be set to either 'inline' or 'import'
    • If programPath is 'inline', arguments are interpreted as script text
    • If programPath is 'import', arguments are interpreted as file path

After using GuestProgramSpec together with an instance of GuestAuthentication as arguments to StartProgramInGuest unique JobID is obtained.

Script progress can be tracked by using the ListProcessesInGuest command. ListProcessesInGuest accepts an array of job id's; passing an empty array will report on all jobs started from API

  • ListProcessesInGuest returns an array of GuestProcessInfo instances:
    • pid field is set to JobID
    • endTime is only set after completion
    • exitCode is set to 0 on success and -1 on error
    • name is set to 'inline' or 'import' (same as programPath in GuestProgramSpec)

Information about completed jobs is kept around for ~1 minute, or untill ListProcessesInGuest (with the corresponding JobID) is called. If the script fails, a file named 'vix_job_$JobID$ .txt' containing the script output is created. Script run time is limited to 120 seconds and script output is not saved on timeout,

  • The vmrun command runScriptInGuest can also be used
  • The PowerCLI cmdlet Invoke-VMScript is not supported
  • Host/guest file transfer is not supported
Python example
#!/usr/bin/env python
# -*- coding: utf-8 -*-

import sys,time
from pyVim import connect
from pyVmomi import vmodl,vim


def runInline(content,vm,creds,source):
    ''' Execute script source on vm '''
    if isinstance(source, list):
        source = '\n'.join(source)
    ps = vim.vm.guest.ProcessManager.ProgramSpec(
                programPath = 'console',
                arguments = source
        )
    return content.guestOperationsManager.processManager.StartProgramInGuest(vm,creds,ps)

def runFromFile(content,vm,creds,fileName):
    ''' Execute script file located on CHR '''
    ps = vim.vm.guest.ProcessManager.ProgramSpec(
                programPath = 'import',
                arguments = fileName
    )
    return content.guestOperationsManager.processManager.StartProgramInGuest(vm,creds,ps)


def findDatastore(content,name):
    sessionManager = content.sessionManager

    dcenterObjView = content.viewManager.CreateContainerView(content.rootFolder, [vim.Datacenter], True)

    datacenter = None
    datastore = None
    for dc in dcenterObjView.view:
        dstoreObjView = content.viewManager.CreateContainerView(dc, [vim.Datastore], True)
        for ds in dstoreObjView:
            if ds.info.name == name:
                datacenter = dc
                datastore = ds
                break
        dstoreObjView.Destroy()

    dcenterObjView.Destroy()

    return datacenter,datastore

def _FAILURE(s,*a):
    print(s.format(*a))
    sys.exit(-1)

#------------------------------------------------------------------------------#

if __name__ == '__main__':
    host = sys.argv[1] # ip or something
    user = 'root'
    pwd = 'MikroTik'
    vmName = 'chr-test'
    dataStoreName = 'datastore1'



    service = connect.SmartConnectNoSSL(host=host,user=user,pwd=pwd)
    if not service:
        _FAILURE("Could not connect to the specified host using specified username and password")

    content = service.RetrieveContent()


    #---------------------------------------------------------------------------
    # Find datacenter and datastore


    datacenter,datastore = findDatastore(content,dataStoreName)

    if not datacenter or not datastore:
        connect.Disconnect(service)
        _FAILURE('Could not find datastore \'{}\'',dataStorename)


    #---------------------------------------------------------------------------
    # Locate vm


    vmxPath = '[{0}] {1}/{1}.vmx'.format(dataStoreName, vmName)
    vm = content.searchIndex.FindByDatastorePath(datacenter, vmxPath)

    if not vm:
        connect.Disconnect(service)
        _FAILURE("Could not locate vm")


    #---------------------------------------------------------------------------
    # Setup credentials from user name and pasword

    creds = vim.vm.guest.NamePasswordAuthentication(username = 'admin', password = '')


    #---------------------------------------------------------------------------
    # Run script

    pm = content.guestOperationsManager.processManager

    try:
        # Run script
        src = [':ip address add address=192.168.0.1/24 interface=ether1;']
        jobID = runInline(content, vm, creds, src)

        # Or run file (from FTP root)
        # jobID = runFromFile(content,vm,creds, 'scripts/provision.rsc')


        #---------------------------------------------------------------------------
        # Wait for job to finish

        pm = content.guestOperationsManager.processManager
        jobInfo = pm.ListProcessesInGuest(vm, creds, [jobID])[0]
        while jobInfo.endTime is None:
            time.sleep(1.0)
            jobInfo = pm.ListProcessesInGuest(vm, creds, [jobID])[0]

        if jobInfo.exitCode != 0:
            _FAILURE('Script failed!')
    except:
        raise
    else:
        connect.Disconnect(service)

Xen

Network, disk, memory and OS info is reported to hypervisor every 30 seconds

  • On older hosts (pre 21.06.2017) only the first ipv4 address assigned to interface is visible

Provisioning

Base-64 encoded strings written to (domain local) xenstore path ('vm-data/provision/script') are decoded and interpreted a script-text. Status ('ready', 'running', 'error') is reported in 'vm-data/provision/script/status'. Scripts are accepted only if 'status' != 'running'. Base 64 encoded script output (if any) is written to 'vm-data/provision/script/otuput'.

  • When creating a VM scripts can be provided by using xenapi (VM.add_to_xenstore_data)

KVM

QEMU guest agent is available. Supported agent commands can be retrieved by using guest-info command. Host-guest file transfer can be performed by using guest-file-* commands. Guest networking information can be retrieved by using the guest-network-get-interfaces command.

  • Scripts can be executed by using the guest-exec command together with the GuestExec data structure:
    • If the path member is provided, the corresponding file is executed
    • If the path member is not set and input-data member is provided, input-data value is used as script input
    • If capture-output is set, script output is reported back
    • args and env members are not used
  • Script job progress can be monitored with guest-exec-status command. The GuestExecStatus data structure is populated as follows:
    • On success exitcode member is set to 0
    • If the script timed out exitcode is set to 1
    • If the script contained errors exitcode is set to -1
    • signal member is not set
    • The err-data member is not used
    • If capture-output was true, Base64 encoded script output is stored in out-data


Proxmox

Some agent commands can be issued by using Proxmox REST api. Guest filesystem quiescing is automatically performed when taking a snapshot

  • Enabling 'chr.provision-agent' for remote access on port 1234

In host shell: 

vmid=256
hostip=192.168.0.1
portnum=1234
qm set $vmid --args "-chardev 'socket,host=$hostip,port=$portnum,id=chr-agent,server,nowait' -device 'virtio-serial,bus=pci.0,addr=0x9' -device 'virtserialport,chardev=chr-agent,name=chr.provision_agent'"
  • Disabling 'chr.provision-agent'

In host shell: 

vmid=256
qm set $vmid --delete args
  • Providing remote access to default agent on port 1234

In host shell: 

vmid=256
portnum=1234
socat TCP-LISTEN:$portnum,reuseaddr,fork UNIX-CLIENT:/run/qemu-server/$vmid.qga

Guest agent python example

import os,time,base64,json,socket,select,errno

class GuestAgent(object):
    '''
        Qemu guest agent interface
        runScript and runFile commands are tailored for ROS agent implementation
        Transport provided by derived classes (transact method)
    '''

    def __init__(self,**kwargs):
        # Due to file contents being passed as base64 inside json:
        #  - large chunk sizes may slow down guest-side parsing.
        #  - small chunk sizes result in additional message fragmentation overhead.
        # Default value is a guestimate.
        self.__chunkSize = kwargs.get('chunkSize', 4096)

    def _qmpError(self,cls,msg):
        ''' Generic callback to log qmp errors before (optionally) raising an exception '''
        print(cls)
        for line in msg.split('\n'):
            print(line)
        # raise RuntimeError()

    def _error(self,msg,*a):
        ''' Generic callback to misc errors before (optionally) raising an exception '''
        print(msg.format(*a))
        # raise RuntimeError()

    def _info(self,msg,*a):
        ''' Generic callback to log info '''
        print(msg.format(*a))

    def _monitorJob(self,pid):
        ''' Block untill script job completes, echo output. Returns None on failure '''
        ret = self.transact('guest-exec-status',{'pid':pid})
        if ret is None:
            return None

        while not bool(ret['exited']):
            time.sleep(1)
            ret = self.transact('guest-exec-status',{'pid':pid})
            if ret is None:
                return None

        # err-data is never sent
        out = []
        if 'out-data' in ret.keys():
            out = base64.b64decode(ret['out-data']).decode('utf-8').split('\n')
            if not out[-1]:
                out = out[:-1]

        exitcode = int(ret['exitcode'])
        return exitcode, out

    def putFile(self,src,dst):
        ''' Upload file '''
        src = os.path.expanduser(src)
        if not os.path.exists(src) or not os.path.isfile(src):
            self._error('File does not exist: \'{}\'', src)
            return None

        ret = self.transact('guest-file-open', {'path':dst,'mode':'w'})
        if ret is None:
            return None

        handle = int(ret)

        file = open(src, 'rb')
        for chunk in iter(lambda : file.read(self.__chunkSize), b''):
            count = len(chunk)
            chunk = base64.b64encode(chunk).decode('ascii')

            ret = self.transact('guest-file-write',{'handle':handle,'buf-b64':chunk,'count':count})
            if ret is None:
                return None
        self.transact('guest-file-flush',{'handle':handle})
        ret = self.transact('guest-file-close',{'handle':handle})
        return True

    def getFile(self,src,dst):
        ''' Download file '''
        dst = os.path.expanduser(dst)

        ret = self.transact('guest-file-open',{'path':src,'mode':'rb'})
        if ret is None:
            return None

        handle = int(ret)
        data = ''
        size = 0

        while True:
            ret = self.transact('guest-file-read',{'handle':handle,'count':self.__chunkSize})
            if ret is None:
                return None
            data += ret['buf-b64']
            size += int(ret['count'])
            if bool(ret['eof']):
                break

        ret = self.transact('guest-file-close',{'handle':handle})
        data = base64.b64decode(data.encode('ascii'))
        with open(dst,'wb') as f:
            f.write(data)
        return True

    def runFile(self,fileName):
        ''' Execute file (on guest) as script '''
        ret = self.transact('guest-exec',{'path':fileName, 'capture-output':True})
        if ret is None:
            return None

        pid = ret['pid']
        return self._monitorJob(pid)

    def runSource(self,cmd):
        ''' Execute script '''
        if isinstance(cmd,list):
            cmd = '\n'.join(cmd)
        cmd += '\n'
        cmd = base64.b64encode(cmd.encode('utf-8')).decode('ascii')

        ret = self.transact('guest-exec',{'input-data':cmd, 'capture-output':True})
        if ret is None:
            return None

        pid = ret['pid']
        return self._monitorJob(pid)

    def shutdown(self,mode='powerdown'):
        '''
            Execut shutdown command
            mode == 'reboot' - reboot guest
            mode == 'shutdown' or mode == 'halt' - shutdown guest
         '''
        ret = self.transact('guest-shutdown',{'mode':mode})
        return ret

class SocketAgent(GuestAgent):
    '''
        GuestAgent using unix/tcp sockets for communication.
    '''
    def __init__(self):
        GuestAgent.__init__(self,chunkSize= 32 * 65536)

    @staticmethod
    def unix(dev):
        ''' Connect using unix socket '''
        self = SocketAgent()
        self.__af = socket.AF_UNIX
        self.__args = dev
        self.__wait = False
        return self

    @staticmethod
    def tcp(ip,port,wait = True):
        ''' Connect using tcp socket '''
        self = SocketAgent()
        self.__af = socket.AF_INET
        self.__args = (ip,port)
        self.__wait = wait
        return self

    def __enter__(self):
        self._sock = socket.socket(self.__af, socket.SOCK_STREAM)
        if self.__wait:
            self._info('Waiting for guest ...')
            # Wait for hyper to create channel
            while True:
                try:
                    self._sock.connect(self.__args)
                    break
                except socket.error as e:
                    if e.errno == errno.EHOSTUNREACH or e.errno == errno.ECONNREFUSED:
                        time.sleep(1)
                    else:
                        self._sock.close()
                        raise

            #Wait for guest agent to initialize and sync
            while True:
                import random
                key = random.randint(0, 0xffffffff)
                msg = json.dumps({'execute':'guest-sync-delimited','arguments':{'id':key}},separators=(',',':'),sort_keys=True)
                self._sock.send(msg.encode('ascii'))

                self._sock.setblocking(0)
                response = b''
                if (select.select([self._sock],[],[])[0]):
                    response += self._sock.recv(65536)
                else:
                    raise RuntimeError()
                self._sock.setblocking(1)

                sentinel = b'\xff'
                response = response.split(sentinel)[-1]
                if not response:
                    time.sleep(3)
                    continue
                response = json.loads(response.decode('utf-8').strip())
                if 'return' in response.keys():
                    if int(response['return']) == key:
                        break
                time.sleep(3)
        else:
            self._sock.connect(self.__args)

        return self

    def __exit__(self,*a):
        self._sock.close()

    def transact(self,cmd,args={}):
        ''' Exchange a single command with guest agent '''
        timeout = 2
        msg = json.dumps({'execute':cmd,'arguments':args},separators=(',',':'),sort_keys=True)
        self._sock.send(msg.encode('ascii'))
        self._sock.setblocking(0)
        response = b''
        if (select.select([self._sock],[],[],timeout)[0]):
            response += self._sock.recv(65536)
        self._sock.setblocking(1)
        if not response:
            response = None
        else:
            if response[0] == 255: # sync
                response = response[1:]
            response = json.loads(response.decode('utf-8').strip())
            if 'error' in response.keys():
                self._qmpError(response['error']['class'],response['error']['desc'])
                response = None
            elif 'return' in response:
                response = response['return']
        return response

#-------------------------------------------------------------------------------

if __name__ == '__main__':
    script = [':log info "hello world";']

    ip = '192.168.0.1'
    port = 1234

    # can also use unix sockets
    #with SocketAgent.unix('/dev/something') as agent:

    with SocketAgent.tcp(ip, port) as agent:
        ret,out = agent.runSource(script)
        print('ret = {}'.format(ret))
        for line in out:
            print(line)


[ Top | Back to Content ]

Retrieved from "https://wiki.mikrotik.com/index.php?title=Manual:CHR&oldid=34287"