I have just finished deploying a new BOA install .. so, version 2.0.8.

I created an initial client/site of open atrium and that succeeded .. but all subsequent attempts to create new site/clients are failing due to this error.

A quick google threw up this link (http://drupal.org/node/951296) .. suggesting that this was a known issue associated with older versions of aegir. So, it does seem like only drupal-7 based platforms would be affected by this, and that explains why the open atrium install worked.

Right now, I am running a find with grep to match all files that have that string, so I can change it to something safer for drupal-7 .. Is there an easier way to fix this?



cnergis’s picture

Attempts to delete the failed site creation are also failing .. due to:

The drush command '@domain.com provision-delete' could not be found.
Could not find a Drupal settings.php file at ./sites/default/settings.php.

So, that begs the question of how to wipe the slate clean so I can recreate the site.

omega8cc’s picture

Status: Active » Closed (works as designed)

There is nothing BOA specific here, and the error on delete attempt of non-existent site is expected - please read the docs.

omega8cc’s picture

Also, please note that will simply close w/o response all issues submitted without required logs attached.

omega8cc’s picture

Category: bug » support
Status: Closed (works as designed) » Closed (cannot reproduce)

The correct status is can't reproduce.

cnergis’s picture

Category: support » bug
Status: Closed (cannot reproduce) » Active

I discovered this link: http://groups.drupal.org/node/57328 .. and followed mig5's advice (after verifying that the platform link was not there (as the failed install had backed it out) .. also verified that the DB had not been created ..

I also discovered the string 'webmaster@localhost' in 4 files and modified them. I finally used the octopus interface to delete the clients.

Now, to run another test site/client provision, and the system refuses to accept the client name claiming it is already in use (even after I had cleared the cache within the octopus interface) ..

Any suggestions on how next to proceed ?

cnergis’s picture

Have just seen your updates above .. so, what logs are suggesting I enclose ?

omega8cc’s picture

Category: bug » support
Status: Active » Closed (cannot reproduce)

Again, this is nothing BOA related. If you believe it is something Aegir related, please submit proper bug report in the Aegir queue, but we can't reproduce this issue on any D7 platform which comes by default with BOA, so it must be something else, I guess.

omega8cc’s picture

cnergis’s picture

Well, I just rebuilt a server (debian squeeze minimal), and even without running any apt updates I installed BOA-208 .. The BOA install (boa in-stable public max) completed successfully but did display an error message about incorrect use of '-s' to mail, and never sent out the email for octopus .. I will check the logs and paste it here. I copied the password reset link to access octopus, and immediately started deploying once all platforms had been verified.

I am mentioning this now in case it is possibly related .. I have no problems with opening a bug, and have just reported this as it occurred .. I did not enclose any logs, because I was unsure which logs would be useful to you guys in diagnosing the problem

cnergis’s picture

unfortunately, it does not seem like octopus keeps install logs .. so, I try to force an update with 'octopus up-stable all both', but this repeatedly fails with the following error:

Octopus [Wed Apr 24 15:48:48 EDT 2013] ==> INFO: Testing GitHub, Drupal and Gitorious servers availability, please wait...
Octopus [Wed Apr 24 15:48:50 EDT 2013] ==> INFO: GitHub mirror repository will be used for this install
Octopus [Wed Apr 24 15:48:50 EDT 2013] ==> UPGRADE in progress...
Octopus [Wed Apr 24 15:48:50 EDT 2013] ==> EXIT: You must enter **your** valid e-mail address in the _CLIENT_EMAIL variable
Octopus [Wed Apr 24 15:48:50 EDT 2013] ==> EXIT: stored both in the /root/.o1.octopus.cnf config file **and**
Octopus [Wed Apr 24 15:48:50 EDT 2013] ==> EXIT: in the /data/disk/o1/log/email.txt file
Octopus [Wed Apr 24 15:48:50 EDT 2013] ==> EXIT: Bye (1)
waiting 4 sec
Done for /data/disk/o1

So, it seems like it is not seeing the variable _CLIENT_EMAIL in /root/.o1.octopus.cnf .. I first wondered if having _MY_EMAIL and _CLIENT_EMAIL the same could have been the issue, but whatever email I set as _CLIENT_EMAIL brings up the same errors .. This could also be related to why the email failed during initial install .. shame I am unable to tell you exactly what the logs showed, but I know for sure it was about '-s' being used to invoke the mail command ..

It is also worth noting that, the .o1.octopus.cnf file I am using is from an old install, and with the placeholders for version and date removed .. Could these be factors ?

I will back off now and try to either fix it, or start all over again or something else .. Don't want to be a bother to you.

Thanks for your responses so far ..

cnergis’s picture

OK .. have made some progress, just as I was about to give up and re-start .. I was just being to hasty in reading the failure output from 'octopus up-stable all both' ..

Specifically, I had missed the continuation ..

Octopus [Wed Apr 24 15:48:50 EDT 2013] ==> EXIT: in the /data/disk/o1/log/email.txt file

A check reveals the /data/disk/o1/log/email.txt file exists but was empty .. so, I added _MY_EMAIL in there, and the upgrade is running now ..

Will carry on from there, and won't bother you again on this subject. Just providing feedback .. thanks again.

cnergis’s picture

The octopus upgrade went fine, and this time I have received an upgrade email ..

Still had issues with re-creating the sites .. but came across the aegir code which throws up that error, and that led me to the right octopus table .. once I delete the remnant entries, I am able to recreate the clients ..

Updating for completeness .. and in hope this may be useful ..