VIFIB DESCENTRALIZED CLOUD COMPUTING

SlapOS is a decentralized Cloud Computing technology that can automate the deployment and configuration of applications in a heterogeneous environment.
1 - 20 of 22 records - 0 items selected
  Summary

Hi there,

 Just started up node and it says  "Bang with message 'Reboot'". In my services under erp5, the instance is not loading and states that its "Looking for a free partition". I've already paid the 2 Euro's and paid shows as received. What am I missing?

Thanks for you help. 

As explained on the FAQ, several things can happen here:

  • you requested an instance of Software Release that is not available (not installed or no place left) in any node you have access to
  • you requested an instance of Software Release on a specific computer, but this computer can't accept this instance (software release not installed or no place left)

Hi,

Normally the payment takes a while to be processed, normally they are processed in batch at midnight only. We are working to improve the speed of check payments. If you didn't do, please open a ticked from your account, the SlapOS master team may help you more.

Regards,

Rafael Monnerat 

From: Unknown User

As explained on the FAQ, several things can happen here:

  • you requested an instance of Software Release that is not available (not installed or no place left) in any node you have access to
  • you requested an instance of Software Release on a specific computer, but this computer can't accept this instance (software release not installed or no place left)

 


Hi,

Am aware that am invoking a very old thread but thought it okay since it is a very similar issue.

I am facing the same problem of 'Looking for a free partition'. Have seen the FAQs - http://community.slapos.org/wiki/slapos-Wiki.Home/slapos-FAQ

I was following How to deploy SlapOS Master instance with SlapOS to install a master on my local network for experimental purposes via the 'slapos console' method. I started the process via slapos.org. Constantly saw the status - 'Looking for a free partition'. Then switched to https://slapos.vifib.com/  . Thereafter, thinking that to have this working I have paid up 2 EUR but yet the invoice is showing as ' Waiting for payment confirmation '.

 Now I have the following questions:

1. What could be the exact reason of the installation not progressing and how can I figure it out? The possible reasons given in the FAQ don't seem to be applicable in my scenario.

2. Is the payment validation still a mid-nightly batch process and should I expect this process to progress after that?

3. Or am I following a no longer supported method to get the SlapOS master running and thereafter ERP5? Please note that my aim is to assess if we can participate in ERP5 development and deployment.

Please do let me know if any further information shall be helpful to guide me.

Any guidance in this regards shall be highly appreciated.

Thanks,

Shrenik

Hi,

Fisrt of all, thank you for using SlapOS platform. The message 'Looking for a free partition' mean that SlapOS Master couldn't allocate your requested instance. The possibles actions and responses for this problem can be found in our FAQ.

About deployment of SlapOS Master, the documentation your are using is outdated, we are updating our developer documentation and we didn't update this documentation yet, please read this documentation for the updated one.

The recommended procedure today to depoy SlapOS Master for development purpose is to request a webrunner then request SlapOS Master inside, this let you easily control your SlapOS Master instance through webrunner. You can follow this documentation to know how to request webrunner on Vifib platform. But you can still to the procedure of SlapOS Master installation in your own computer like your started before.

After you have requested your webrunner, go to "Menu" -> " Open Software Release" then select 'slapos-master' in software folder and open it. You can now click on the green button and inspect the result.

About questions:

1- If you still have the message 'Looking for a free partition' after you have paid the subscription invoice of 2 €, this can mean than your Software Release is not installed into the computer your are trying to use. You can request installation of Software Release from slapconsole with Supply command like in the documentation you are following for SlapOS Master deployment. Make sure that you have used the same computer Id (COMP-XXX) in your supply and request command.

If you compile software release from your own computer, you can also check log in folder /opt/slapos/log/slapos-node-software.log. If your instance is under deployment and is reporting errors, you can check log in : /opt/slapos/log/slapos-node-instance.log

2-  Normally the payment validation should not be so long. Please let us know how long it take to validate your payment so we can improve if it's needed. You can also create a tiket in slapos.org / slapos.vifib.com so we will have more information to correctly handle the problem.

3- It's great that you can  participate in ERP5 development and deployment, please try above procedure for SlapOS Master deployment and let us know if you have any issue.

Regards,

Alain Takoudjou 

Hi Alain,

Thanks a lot for your response.

I am trying the non-webrunner method first from http://community.slapos.org/wiki/developer-Allocate.SlapOS.Master.cluster.Instance. But there seems to be something wrong.

This was my supply command-

 

software_url = 'https://lab.nexedi.com/nexedi/slapos/raw/master/software/slapos-master/software.cfg'
supply(
    software_url,
    'COMP-2547'
)

This was my request command -

import json
software_url = 'https://lab.nexedi.com/nexedi/slapos/raw/master/software/slapos-master/software.cfg'
# Choose a SlapOS Node
# If you are deploying SlapOS Master with Webrunner, then computer_id is 'slaprunner'
computer_id = 'COMP-2547'

parameter_dict = {
  "timezone": "UTC",
  "site-id": "erp5",
  "bt5": "erp5_full_text_myisam_catalog slapos_configurator",
  "zope-partition-dict": {
    "activities-node": {
      "family": "activities",
      "thread-amount": 3,
      "instance-count": 1,
      "timerserver-interval": 1,
      "computer-guid": computer_id,
    },

    "distribution-node": {
      "family": "distribution",
      "thread-amount": 1,
      "instance-count": 1,
      "computer-guid": computer_id,
      "port-base": 2210,
      "timerserver-interval": 1,
    },

    "admin-node": {
      "family": "admin",
      "thread-amount": 1,
      "instance-count": 1,
      "port-base": 2220,
      "computer-guid": computer_id
    },
    
    "web-node": {
      "family": "web",
      "thread-amount": 2,
      "instance-count": 1,
      "computer-guid": computer_id,
      "port-base": 2230,
    },

    "service-slapos": {
      "family": "service",
      "thread-amount": 2,
      "instance-count": 1,
      "computer-guid": computer_id,
      "port-base": 2240,
      "ssl-authentication": True,
      "backend-path": "/%(site-id)s/portal_slap",
    },
  },

  "mariadb-computer-guid": computer_id,
  "mariadb": {
    "test-database-amount": 0
  },

  "zodb-software-type": "zeo",
  "zodb-computer-guid": computer_id
}

# If your are deploying SlapOS Master with Webrunner, don't set "certificate-authority-path" parameter
ca_path = '/etc/slapos-ssl'
parameter_dict["certificate-authority-path"] =  ca_path

# Choose a title
title = "SlapOS Master on COMP-2547"

request(software_url,
  title,
  filter_kw={'computer_guid': computer_id},
  software_type='create-erp5-site',
  partition_parameter_kw={
   '_': json.dumps(parameter_dict, sort_keys=True, indent=2),
  }
)

But I see this in Home > My Space > My Services -> clicking on the service - #error while instanciating

Error while instansiating

 

 

And as you guided I have checked the log at /opt/slapos/log/slapos-node-instance.log which has this error which I have highlighted in red -

[2016-03-15 18:12:03,748] DEBUG    Loading config: /etc/opt/slapos/slapos.cfg
[2016-03-15 18:12:03,766] INFO     Sleeping for 11 seconds. To disable this feature, check --now parameter in slapgrid help.
[2016-03-15 18:12:16,325] INFO     Processing computer partitions...
[2016-03-15 18:12:16,335] DEBUG    Supervisord already running.
[2016-03-15 18:12:18,270] DEBUG    Check if slappart19 requires processing...
[2016-03-15 18:12:18,271] DEBUG    Partition already up-to-date, skipping.
[2016-03-15 18:12:18,272] DEBUG    Check if slappart18 requires processing...
[2016-03-15 18:12:18,286] INFO     Processing Computer Partition slappart18.
[2016-03-15 18:12:18,287] INFO       Software URL: https://lab.nexedi.com/nexedi/slapos/raw/master/software/slapos-master/software.cfg
[2016-03-15 18:12:18,288] INFO       Software path: /opt/slapgrid/bcb8fbfbae583c14297f14ec8ebc3724
[2016-03-15 18:12:18,288] INFO       Instance path: /srv/slapgrid/slappart18
[2016-03-15 18:12:20,702] INFO       Instance status: started
[2016-03-15 18:12:20,704] INFO     Installing Computer Partition slappart18...
[2016-03-15 18:12:20,712] DEBUG    Overridden HOME = '/srv/slapgrid/slappart18'
[2016-03-15 18:12:20,713] DEBUG    Removed from environment: LANG, LOGNAME
[2016-03-15 18:12:20,714] ERROR
Traceback (most recent call last):
  File "/opt/slapos/eggs/slapos.core-1.3.15-py2.7.egg/slapos/grid/slapgrid.py", line 1153, in processComputerPartitionList
    self.processComputerPartition(computer_partition)
  File "/opt/slapos/eggs/slapos.core-1.3.15-py2.7.egg/slapos/grid/slapgrid.py", line 1027, in processComputerPartition
    local_partition.install()
  File "/opt/slapos/eggs/slapos.core-1.3.15-py2.7.egg/slapos/grid/SlapObject.py", line 517, in install
    self.software_release_url, template_location))
IOError: Software Release https://lab.nexedi.com/nexedi/slapos/raw/master/software/slapos-master/software.cfg is not correctly installed.
Missing file: /opt/slapgrid/bcb8fbfbae583c14297f14ec8ebc3724/instance.cfg
[2016-03-15 18:12:22,312] INFO     Finished computer partitions.

1. Can you please tell me what have I done incorrectly?

2. The payment issue is resolved and I see the invoice is now marked as Paid, hence I assume the account is now validated.

3. If I deploy my own slapos master, will I still be charged an amount of 2 EUR on a daily basis?

Look forward to your guidance.

Thanks,

Shrenik




And the log is filling up with the same message every minute. What do I need to terminate/clean-up and restart the process?

As per the discussion at http://community.slapos.org/forum/Missing-file-instance.cfg,  I don't yet have the .completed file in /opt/slapgrid/bcb8fbfbae583c14297f14ec8ebc3724

However, I see that around 5 GB of files have been downloaded/created.

~# du -s /opt/slapgrid/bcb8fbfbae583c14297f14ec8ebc3724/*
200     /opt/slapgrid/bcb8fbfbae583c14297f14ec8ebc3724/bin
4       /opt/slapgrid/bcb8fbfbae583c14297f14ec8ebc3724/buildout.cfg
25500   /opt/slapgrid/bcb8fbfbae583c14297f14ec8ebc3724/develop-eggs
92008   /opt/slapgrid/bcb8fbfbae583c14297f14ec8ebc3724/eggs
5012148 /opt/slapgrid/bcb8fbfbae583c14297f14ec8ebc3724/parts

So some action is happening but I am unable to make sense of it. Am not sure if I should do anything or simply wait till the .completed file is generated.

Any guidance on what I should do next shall be helpful.

 Hi,

1- About error message bellow:

IOError: Software Release https://lab.nexedi.com/nexedi/slapos/raw/master/software/slapos-master/software.cfg is not correctly installed.
Missing file: /opt/slapgrid/bcb8fbfbae583c14297f14ec8ebc3724/instance.cfg

It mean that the software release is not correctly compiled. If compilation of sotware release fail, SlapOS will retry until it succeed. I think all you did was ok, but the SlapOS Master Software doesn't compile in your distro, it need to be fixed. In your case, you can get compilation error message with: tail -f /opt/slapos/log/slapos-node-software.log

Although SlapOS will retry to compile your software release to fix errors, if some file cannot be downloaded for example, the compilation will always fail. So you must check the log to make sure that Software can compile. Can you show what you have in your /opt/slapos/log/slapos-node-software.log ?

2- Yes, your instance is now start requested this mean that your account has been validated.

3- As your are deploying instance in your own server, you will only need to pay 1€ per month for instance subscription, if you have 2 instances, it will be 2 €, ...

Regards,

Alain Takoudjou

 

Hi Alain,

I have uploaded the current log herein - https://www.dropbox.com/s/mvndbz47bzz58e5/slapos-node-software.rar?dl=0

Please check and let me know.

Thanks,

Shrenik


 Hi,

SlapOS is failing to build OpenBLAS in your distro. We mostly test software releases on debian and CentoOS distribution. Which one are you using ?

Here is the error message:

Makefile.system:393: *** OpenBLAS: Cannot set both USE_OPENMP=1 and USE_THREAD=0. The USE_THREAD=0 is only for building single thread version.. Stop.
[2016-03-16 20:17:08,116] INFO openblas: Command failed with exit code 2: make BINARY="$(uname -m | grep -q x86_64 && echo 64 || echo 32)" NO_STATIC=1 USE_OPENMP=1 USE_THREAD=1 || (make -j1 clean && make BINARY="$(uname -m | grep -q x86_64 && echo 64 || echo 32)" NO_STATIC=1 USE_OPENMP=1 USE_THREAD=1 TARGET=GENERIC)
[2016-03-16 20:17:08,116] INFO openblas: Compilation error. The package is left as is at /opt/slapgrid/bcb8fbfbae583c14297f14ec8ebc3724/parts/openblas__compile__/OpenBLAS-0.2.15 where you can inspect what went wrong

 

You are compiling from this version of openblas component: https://lab.nexedi.com/nexedi/slapos/blob/master/component/openblas/buildout.cfg

Are you using a debian machine ? so I can try to reproduce the error.

Regards,

Alain Takoudjou

Hi Shrenik,

I checked OpenBLAS's Makefile.system and found that USE_THREAD is overridden to 0 if it is built on a single core machine. I just added a patch so that USE_OPENMP should be also overridden to 0 in such case.

https://lab.nexedi.com/nexedi/slapos/commit/73f9574f4a9868d468693432575330231f750353 

I am using Ubuntu 14.04. Any suggestions to fix this on this OS itself?

Hi Kazuhiko,

Could you please guide me what I need to do with this patch to move ahead?

Do I need to manually apply it to the OpenBLAS module code and thereafter wait while the compilation auto-fixes?

Okay, I applied the changes as shared in the commit and the compilation seems to be progressing ahead for now.

Thanks,

Shrenik

Hi,

Now I have a .completed file in /opt/slapgrid/bcb8fbfbae583c14297f14ec8ebc3724 . But I am not sure what to do next as it is not clear from this document - http://community.slapos.org/wiki/developer-Allocate.SlapOS.Master.cluster.Instance . The last step I reached was the request command.

I thought the next step would be "Fix Certificate Authority path (Only if you are not deploying in Webrunner)" but it doesn't seem possible as the .pem file is missing from the dir /etc/slapos-ssl .

Even I see this image I see a few reds -

Service Status

Please help me make sense of this.

Thanks,

Shrenik

Hi Shrenik,

Balancer instance is faling to deploy. Can you check what you have in /opt/slapos/log/slapos-node-instance.log ?

Regards,

Alain Takoudjou

 

Hi Alain,

It's odd but I don't spot anything obvious that is wrong in the log.

Herein is a link to the log archive - https://dl.dropboxusercontent.com/u/73028350/slapos-node-instance.rar

I thought that this entire process over slapos was supposed to be very simplified but it seems no one seems to be using it this way and hence it is all unmaintained, therefore all these hiccups in deployment. Is it true that webrunner and online instance at vifib is the only way that all developers are working now?

Do let me know if you want me to try anything else.

Thanks,

Shrenik

 hi Shrenik,

By checking your log file, I realized that there was a mistake in request parameters. We did a commit recently to improve the software release but the documentation was not updated properly. Sorry, this is just a problem of tipo.

In your request parameters, section "REQUEST SLAPOS MASTER INSTANCE" please change this:

parameter_dict["certificate-authority-path"] = ca_path   by   parameter_dict["shared-certificate-authority-path"] = ca_path and request your instance like you did at the first time to update parameters.

If you want, you can also edit that parameter from vifib.com / slapos.org  interface. Go to you instance, search "certificate-authority-path" in your instance XML and replace by "shared-certificate-authority-path". Finally, click on "Update XML" to apply changes.

By applying the changes, your instance will be reprocessed but not all important processes will be restarted. Stop your instance and wait about 5 minutes then start again your instance. The problem will be fixed.

Check that "certificate authority folder" is not empty and that it contain the file "cacert.pem". When you confirm that all are OK, you can continue the documentation at the point "FIX CERTIFICATE AUTHORITY PATH".

I modified the documentation to fix the parameter name.

Please continue your deployment like you started to do. If you want, you can try later the deployment with webrunner, but both cases should work we are doing our best to fix all possibles issues. By using it, you help us to improve the software and the documentation too.

Thank you,

Alain Takoudjou

 

 

Thank you Alain,

I have made some progress. But have a few observations and queries, especially around the documentation.

>1. "service-v6": It's the url that should be used as master_url in all SlapOS Nodes configuration, generally in /etc/opt/slapos/slapos.cfg.

So should I change 'master_url = https://slap.vifib.com' in  /etc/opt/slapos/slapos.cfg to the service-v6 value - https://[2001:67c:1254:12::83b2]:2157 ?


>Deploy Frontend

>To access to the web interface of the SlapOS Master, you will need to setup a reverse proxy (a.k.a "Apache frontend") with slapos. By default, it will use infrastructure of slapos.org to act as a frontend. But you can use yours.

So I assume that this step is optional and I can continue using slapos.org as the frontend for the master. Is my understanding correct? Is the process to configure an Apache based reverse proxy documented with any specifics that may be required in this scenario?

> Request Slave Frontend Instance
> From the SlapOS console: type the following command. You should replace 'COMP-XXXX' in the end by the identifier of your SlapOS Node:

The instructions should be to replace both COMP-XXXX and WEB-v6 URL.

I used this:

request(
  'http://git.erp5.org/gitweb/slapos.git/blob_plain/HEAD:/software/apache-frontend/software.cfg',
  'Slave frontend for SlapOS Master on COMP-2547',
  partition_parameter_kw={
    'url': 'https://[2001:67c:1254:12::83b2]:2159',
    'type': 'zope',
    'path': 'erp5',
  },
  shared=True,
)

I was able to login but upon first login encountered this error. See screenshot.



With these instructions it seems that the resulting erp5 website is being hosted at https://softinst65419.host.vifib.net which means that it is not on my local server. Shouldn't the "SlapOS Master Slave frontend" be also hosted by my local server or is it? Sorry, I am completely lost on this concept.

>Site Consistency and Configurator

>    First solve Site Consistency like described at:
>    http://www.osoe-project.org/user-Howto.Manage.Promise.On.ERP5/view

The link returns a 404. Hence I continued with some guess work and clicked on 'Check Site Consistency' from the 'My Favourites' menu.

There were some issues -


 I chose 1st and 4th and hit 'Force Reload Promises' and 2nd and 3rd and hit 'Fix Site Configuration'. I hope I have done the right thing. All this was guesswork as I am growing impatient with this deployment and my development team is also waiting far too long on this.

>Check /portal_activities/manageActivities and wait until there is no activity left. You can periodically reload the page to see the status.

There are no activities listed. At least I never could see any whenever I have opened this page.

>Solve again consistency if any consistency error is still present and wait for background activities to finish.

No error to handle.

>As soon "Configure Your Site" is Showed on menu (and make sure you don't have activities) you can guide yourself with :
http://www.osoe-project.org/user-Howto.Configure.ERP5.for.SMB.With.Configurator/view

The only option I get is this -



I clicked on Configure and then went back to check something and then clicked on Configure again and I did not get any options to configure but the installation started and the URL is this -

http://softinst65419.host.vifib.net/portal_configurator/login?field_my_user_preferred_language=en&dialog_id=ConfiguratorTool_view&default_field_my_user_preferred_language:int=0&dialog_method=login&field_your_business_configuration=business_configuration_module/slapos_master_configuration_workflow&update_method=&cancel_url=http%3A//softinst65419.host.vifib.net/portal_configurator/ConfiguratorTool_view&user_preferred_language=en&dialog_category=None&form_id=view

Do let me know if you need any further information. I am trying to proceed ahead from here. It has been around 20 minutes and the screen is showing this, but is progressing slowly -



and there is a lot of activity/load on the server. Server is a VM with 4 cores and 8 GB memory. I am hoping this will complete soon.


I hope this feedback is good enough to make certain improvements to the documentation. Look forward to your thoughts on my observations above.

Is it possible that we can communicate over some IRC or IM such as Skype or Google hangout, etc. because this communication may not actually benefit anyone since most are issues that will happen again once the documentation is perfected and if we interact better, then we may be able to test faster and improve the documentation even faster?

Regards,S
Shrenik