or Create a profile

Home → Problems →

rackspace server instances terminated

bluethundr's Avatar

bluethundr

03 Feb, 2012 06:07 PM via web

Hello,

I've just added rackspace cloud to my scalr account. One thing that I've noticed is that the instances that I launch from scalr keep being terminated with no intervention from me. This is on the JF-RACKCOULD farm. I am the only user of this scalr account. How could this be happening and how can we make this stop? :)

Also not sure if you can help here but I am happily using the scalr comand line tools with my AWS account. I would like to be able to use the scalr commands with both the AWS account as well as the Rackspace account. I did try to run scalr configure from the command line and entered my API id and API access key from the RackSpace side of things. But when I do a scalr list farms command all I see is the AWS side of things.

I'd appreciate any input you may have on both of these problems..

Thanks
tim

  1. 2 Posted by bluethundr on 03 Feb, 2012 07:32 PM

    bluethundr's Avatar

    also it appears that a rackspace instance I started with scalr is running but scalr is not aware of it.

    50.56.193.205

    I know this by the fact that this rackspace instance already has scalarizr installed and running. thoughts/suggestions here?

    tx

  2. 3 Posted by bluethundr on 04 Feb, 2012 04:36 PM

    bluethundr's Avatar

    also please note that this problem only occurs with instances launched from scalr. If I launch from within the cloud control panel on rackspace this does not happen. And it appears that the instances are NOT actually terminated in rackspace although scalr is reporting that they are. I hate to say this but from my perspective using scalr with rackspace cloud is significantly buggier than using it with AWS which has generally been a smooth experience.

    thank you
    tim

  3. Support Staff 4 Posted by Igor Savchenko on 04 Feb, 2012 08:18 PM

    Igor Savchenko's Avatar

    I cannot find this farm under your account. If server is running at Rackspace and Pending at Scalr - it's mean that something wrong with communication between scalr and instance or maybe it's an issue with scalarizr on the instance. Can you please create test farm with rackspace role (which doesn't work) and let us start it and check what is wrong.

    Thanks,
    Igor

  4. 5 Posted by bluethundr on 05 Feb, 2012 10:57 PM

    bluethundr's Avatar

    Hi Igor,

    Thanks for looking into this. The farm is there it is ID 8371. Sorry I didn't include the farm id in the original correspondence.

    Not sure why you didn't see this but it has been there since I've setup rackspace to work with scalr a few days ago. :)

    I've also tried launching a server instance a little while ago under the rackspace side of things and the behavior was the same. I launched a server under farm 8371 went out for a few hours and when I came back I logged back into scalr and rackspace cloud. There was no evidence of the instance in the scalr interface or in rackspace cloud.

    At 6:05 pm (which was right for the time that I launched the instance through the Scalr interface) I got this email form rackspace cloud:

    Your Cloud Server build is complete.

    IP: 50.56.199.152
    root/Administrator password: 971a0392-31ab-491a-bb7a-e88947644b94kRl1M67Lr

    Well I don't really approve of rackspace cloud sending the root password in an email upon launching instances. However if you're on top of it and know what to expect I guess it isn't all that bad. Still I wish they'd handle things the way AWS does. End of rant.

    At least in the case of 50.56.194.22 could you please delete it from the scalr database? Scalr is not showing this server under my racskspace account and when I tried to re-import it claimed that it couldn't because that IP was already in the scalr database. At least in the case of this one instance it appears that scalr reported that the instance was terminated while it kept running in the rackspace cloud. Usually however after launching rackspace instances they are terminated both in rackspace cloud and in scalr with no intervention with me. I do not have this problem in using Scalr with my AWS account.

    Thanks
    Tim

  5. Support Staff 6 Posted by Marat Komarov on 06 Feb, 2012 06:22 PM

    Marat Komarov's Avatar

    This role has MySQL 5.5 which is currently unsupported by Scalarizr. We are working on 5.5 support and will release it on Scalr.net next month

  6. 7 Posted by bluethundr on 06 Feb, 2012 09:19 PM

    bluethundr's Avatar

    Cool thanks for the info...I look forward to working with rackspace and scalr around then. Am I correct in understanding that there is no way to reimport the orphaned instance I have running at rackspace in the meantime?

    This is it's ip: 50.56.194.22

  7. Support Staff 8 Posted by Marat Komarov on 07 Feb, 2012 09:04 AM

    Marat Komarov's Avatar

    You can of course. Cancel previous server (if you have it in Importing state) and import server again. This will create another new role.

  8. 9 Posted by bluethundr on 07 Feb, 2012 01:48 PM

    bluethundr's Avatar

    Thanks,

    This is what happens when I try to re-import the node in rackspace :

    [root@jf-rackcloud ~]# scalarizr --import -y -o server-id=6a21df05-d5fa-4e04-ad18-be3fe38791a7 -o role-name=JF-DEV-RACK -o crypto-key=ejkgF7PeRFj92eeEfVNFL0hXQ0YTz+oYDL8HGLSjrjbdMXkIjxBIUg== -o platform=rackspace -o behaviour= -o queryenv-url=https://my.scalr.net/query-env -o messaging-p2p.producer-url=https://my.scalr.net/messaging 2012-02-07 13:44:33,886 - INFO - scalarizr - [pid: 1602] Starting scalarizr 0.7.172
    Starting import process...
    Don't terminate Scalarizr until Scalr will create the new role
    2012-02-07 13:44:33,887 - INFO - scalarizr.config - State: importing
    2012-02-07 13:44:34,121 - ERROR - scalarizr - No module named mysql
    Traceback (most recent call last):
    File "/usr/lib/python2.6/site-packages/scalarizr/init.py", line 645, in main

    cnf.reconfigure(values=CmdLineIni.to_kvals(optparser.values.cnf), silent=True, yesall=True)

    File "/usr/lib/python2.6/site-packages/scalarizr/config.py", line 822, in reconfigure

    return self._reconfigure(values, silent, yesall, nodefault, onerror, dryrun)

    File "/usr/lib/python2.6/site-packages/scalarizr/config.py", line 570, in call

    self._configure_handler(name, ini.get('handlers', name), kvals, silent, yesall, nodefault, onerror, dryrun)

    File "/usr/lib/python2.6/site-packages/scalarizr/config.py", line 508, in _configure_handler

    __import__(module_name)

    File "/usr/lib/python2.6/site-packages/scalarizr/handlers/ip_list_builder.py", line 13, in

    from scalarizr.handlers.mysql import MysqlMessages

    ImportError: No module named mysql

  9. Support Staff 10 Posted by Marat Komarov on 07 Feb, 2012 04:00 PM

    Marat Komarov's Avatar

    Scalarizr package is broken on your server. I have no idea why this may happen.
    Reinstall it with:

    $ yum erase scalarizr-rackspace scalarizr-base

    $ yum install scalarizr-rackspace

    Can you can provide steps to reproduce this issue?

  10. Marat Komarov closed this discussion on 01 Mar, 2012 09:47 AM.

  11. bluethundr re-opened this discussion on 10 Mar, 2012 08:08 PM

  12. 11 Posted by bluethundr on 10 Mar, 2012 08:08 PM

    bluethundr's Avatar

    Hi Marat,

    Sorry I never got back to you on this issue. I've been focusing on working with AWS in the hopes that this problem would be resolved on the scalr side of things. But the problem is still there.

    The steps to reproducing this are as follows:

    1) go to the upper right where there is the circular arrows and it says 'default' and click. Select 'RackSpace Cloud' from the drop-down 2) Server Farms -> JF-RACSKPACE -> Roles
    3) JF-RACK -> Options -> Launch New Instance

    That will launch an instance just as it would on the AWS side. But after about 10 or 15 minutes that new instance will show as 'terminated' in Scalr, but continue running in RackSpace cloud. Last time I tried this (about a month or so ago) was able to re-import the instance but then the same thing would happen. Scalr would forget the instance. It's very annoying!

    I would really like to use scalr with rackspace cloud as it's much cheaper than AWS. But I can't work with it until this issue is resolved.

    Thanks
    tim

  13. 12 Posted by bluethundr on 10 Mar, 2012 09:33 PM

    bluethundr's Avatar

    Also I don't see any roles available to me on the Racskspace side. If I go to add a role under my rackspace farm all I see is the role that I created a while ago. Shouldn't there be a number of roles available for rackspace?

  14. Support Staff 13 Posted by Nick Toursky on 10 Mar, 2012 10:20 PM

    Nick Toursky's Avatar

    Regarding the latter question.
    Since Rackspace doesn't support image publishing, you need to build a role first. Use Roles->Build new.

  15. Support Staff 14 Posted by Marat Komarov on 12 Mar, 2012 05:46 PM

    Marat Komarov's Avatar

    The problem with JF-RACK role: that is bundled with MySQL 5.5 which is currently unsupported by Scalarizr. MySQL 5.1 is the highest supported version.

  16. 15 Posted by bluethundr on 15 Mar, 2012 01:13 PM

    bluethundr's Avatar

    ok I'll give that a try! thanks!

  17. Marat Komarov closed this discussion on 12 Apr, 2012 09:41 AM.

Comments are currently closed for this discussion. You can start a new one.

4 people watching.

  • New Issue

  • Conversation Started

  • The discussion is closed

    No more actions from Scalr.net or the discussion starter are required.

  • Re-open the discussion

Recent Discussions

21 May, 2012 06:16 PM Rackspace launching farm
21 May, 2012 01:35 PM Additional servers fail to launch
21 May, 2012 01:30 PM Problem migrating server using scalarizer

Recent Articles

SNMP Restart
List of Scalr API methods
Using the Scalr API
Your own Scalr using Open Source code
Getting Started