Hi,

I recently upgraded BOA installation from 2.0.3 to 2.0.8 on my VPS.

At first I had some problems with the whole upgrade-> MySQL access denied. After a search or two I found one case where they suggested that password had been manually changed. So I went and took password from .my.conf. and updated db tables.
Now the upgrade went trough and during the installation there was no errors or such.
However, after upgrading barracuda and octopus I got 404 on aegir front-end and 500 on site(octopus).

I reviewed nginx access log and it said 13: permission denied (when accessing aegir frontend).
Therefore I made some changes to file permission on var/aegir/host_master with reference to old installation(still in the folder).
*And now I can access Aegir frontend though update doesn't work.
What would be the correct file permissions in different data sources? Same as Drupal? Is there any script available that could take care of fixing file permissions?

I can also receive emails about available updates concerning my current installation.
Should I just try to install it one more time with the debug mode set to true?

*Edits and updates

  • Browsing to VPS IP returns Chive login and so does chive.o1.dot.com
  • All other subdomains except for aegir o1/front-end =404 do return nginx default under construction page.
  • I reviewed MySQL database permissions and found out that there were only a few db permissions. Therefore I had to create couple new db permissions and the dot.com started to work. But still, hostmaster/aegir front-end doesn't open (404).

Comments

FYI
I just tried to upgrade it again with debug mode set to yes and barracuda up-stable log command.

And the result was

BARRACUDA upgrade complete
BYE!

And I'm back to square one: hostmaster(aegir) -> 404 and site ->500.

StatusFileSize
new125.67 KB

I found the latest barracuda log(attachment).

Project:Barracuda» Octopus
Issue summary:View changes

edits and updates

Issue summary:View changes

edits and updates

Issue summary:View changes

edits and updates

Project:Octopus» Barracuda
Status:Active» Postponed (maintainer needs more info)

The system and Master Instance upgrade looks fine, no errors there.

Could you also enable Drush debugging in the Octopus config file /root/.USER.octopus.cnf and post the full debug output on another upgrade attempt with octopus up-stable all log - it will be written to /var/backups/reports/up/octopus/DATE/octopus-up-* file.

Project:Barracuda» Octopus
Issue summary:View changes

edits and updates

Issue summary:View changes

edits and updates, clean up

Project:Octopus» Barracuda
StatusFileSize
new1.96 KB

Let's take one step backwards and review the log file related to first upgrade attempt on octopus.

Don't know how I didn't see this earlier.

Octopus [ke 17.4.2013 22.48.59 +0300] ==> BOA Skynet welcomes you aboard!
Octopus [ke 17.4.2013 22.49.02 +0300] ==> INFO: Reading your /root/.o1.octopus.cnf config file
Octopus [ke 17.4.2013 22.49.03 +0300] ==> NOTE! Please review all config options displayed below
###
### Configuration created on 121013-1951 with
### Octopus version BOA-2.0.3
###
### NOTE: the group of settings displayed bellow
### will *override* all listed settings in the Octopus script.
###
_USER="o1"
_MY_EMAIL="root@localhost"
_PLATFORMS_LIST="ALL"
_AUTOPILOT=NO
_HM_ONLY=NO
_O_CONTRIB_UP=YES
_DEBUG_MODE=NO
_MY_OWNIP=
_FORCE_GIT_MIRROR=""
_THIS_DB_HOST=localhost
_DNS_SETUP_TEST=NO
_HOT_SAUCE=NO
_USE_CURRENT=YES
_REMOTE_CACHE_IP=127.0.0.1
_LOCAL_NETWORK_IP=
_PHP_FPM_VERSION=5.3
_PHP_CLI_VERSION=5.3
###
### NOTE: the group of settings displayed bellow will be *overriden*
### by config files stored in the /data/disk/o1/log/ directory,
### but only on upgrade.
###
_DOMAIN="o1.hostname"
_CLIENT_EMAIL="root@localhost"
_CLIENT_OPTION="CLASSIC"
_CLIENT_SUBSCR="M"
_CLIENT_CORES="4"
###
### Configuration created on 121013-1951 with
### Octopus version BOA-2.0.3
###
_ALLOW_UNSUPPORTED=NO
_USE_STOCK=NO
_STRONG_PASSWORDS=NO
Do you want to proceed with the upgrade? [Y/n] /var/backups/OCTOPUS.sh.o1: line 589: read: read error: 0: Input/output error
Octopus [to 18.4.2013 00.20.21 +0300] ==> INFO: Testing GitHub, Drupal and Gitorious servers availability, please wait...
Octopus [to 18.4.2013 00.20.23 +0300] ==> INFO: GitHub mirror repository will be used for this install
Octopus [to 18.4.2013 00.20.24 +0300] ==> UPGRADE in progress...
Octopus [to 18.4.2013 00.20.24 +0300] ==> START -> checkpoint:
  * Your Aegir control panel for this instance is available at https://o1.hostname
  * Your Aegir system user for this instance is o1
  * This Octopus will use PHP-CLI 5.3 for all sites
  * This Octopus will use PHP-FPM 5.3 both for D6 and D7 sites
  * This Octopus includes platforms: ALL / Unsupported: NO
  * This Octopus options are listed as CLASSIC / M / 4 C
Do you want to proceed? [Y/n] Octopus [to 18.4.2013 00.20.24 +0300] ==> 8s before we will continue...
Octopus [to 18.4.2013 00.20.34 +0300] ==> INIT A: Aegir automated install script part A
Octopus [to 18.4.2013 00.20.34 +0300] ==> INFO A: Checking OCTOPUS version, please wait...
Octopus [to 18.4.2013 00.20.34 +0300] ==> INFO: Version test result: OK
Octopus [to 18.4.2013 00.20.34 +0300] ==> UPGRADE A: UPGRADE
Octopus [to 18.4.2013 00.20.36 +0300] ==> UPGRADE A: Shared platforms code v.002 (latest available) will be used for this install
Octopus [to 18.4.2013 00.20.36 +0300] ==> UPGRADE A: Creating directories with correct permissions, please wait...
Octopus [to 18.4.2013 00.20.38 +0300] ==> UPGRADE A: Syncing provision backend db_passwd, please wait...
Octopus [to 18.4.2013 00.20.42 +0300] ==> UPGRADE A: Running hosting-dispatch (1/3), please wait...
Octopus [to 18.4.2013 00.20.49 +0300] ==> UPGRADE A: Running hosting-dispatch (2/3), please wait...
Octopus [to 18.4.2013 00.20.56 +0300] ==> UPGRADE A: Running hosting-dispatch (3/3), please wait...
Do you want to upgrade this Aegir Satellite Instance? [Y/n] Octopus [to 18.4.2013 00.21.00 +0300] ==> UPGRADE A: Syncing hostmaster frontend db_passwd, please wait...
Octopus [to 18.4.2013 00.21.02 +0300] ==> UPGRADE A: Switching user and running AegirSetupB, please wait...
Octopus [to 18.4.2013 00.21.04 +0300] ==> INIT B: Aegir automated install script part B
Octopus [to 18.4.2013 00.21.04 +0300] ==> UPGRADE B: Creating directories with correct permissions
Octopus [to 18.4.2013 00.21.06 +0300] ==> UPGRADE B: Running standard installer
Octopus [to 18.4.2013 00.21.08 +0300] ==> UPGRADE B: Downloading drush, please wait...
Octopus [to 18.4.2013 00.21.10 +0300] ==> UPGRADE B: Drush seems to be functioning properly
Octopus [to 18.4.2013 00.21.10 +0300] ==> UPGRADE B: Installing provision backend in /data/disk/o1/.drush
Octopus [to 18.4.2013 00.21.10 +0300] ==> UPGRADE B: Downloading Drush and Provision extensions, please wait...
Octopus [to 18.4.2013 00.21.13 +0300] ==> UPGRADE B: Testing previous install...
Octopus [to 18.4.2013 00.21.13 +0300] ==> UPGRADE B: Hostmaster STATUS: upgrade start
Octopus [to 18.4.2013 00.21.15 +0300] ==> UPGRADE B: Running hostmaster-migrate, please wait...
Octopus [to 18.4.2013 00.21.41 +0300] ==> UPGRADE B: Hostmaster STATUS: upgrade completed
Octopus [to 18.4.2013 00.21.41 +0300] ==> UPGRADE B: Simple check if Aegir upgrade is successful
Octopus [to 18.4.2013 00.21.43 +0300] ==> UPGRADE B: FATAL ERROR: Required file /data/disk/o1/aegir/distro/003/sites/o1.hostname/settings.php does not exist
Octopus [to 18.4.2013 00.21.43 +0300] ==> UPGRADE B: FATAL ERROR: Aborting AegirSetupB installer NOW!
Octopus [to 18.4.2013 00.21.43 +0300] ==> UPGRADE A: FATAL ERROR: AegirSetupB installer failed
Octopus [to 18.4.2013 00.21.43 +0300] ==> UPGRADE A: FATAL ERROR: Aborting AegirSetupA installer NOW!
Octopus [to 18.4.2013 00.21.43 +0300] ==> FATAL ERROR: AegirSetupA installer failed
Octopus [to 18.4.2013 00.21.43 +0300] ==> FATAL ERROR: Aborting Octopus installer NOW!

Ok, thanks for reviewing.

I've now tried to upgrade octopus according to your guidance. Here's what happened:
What concerns me is that
instance control panel is available at o1.MY_HOSTNAME since it should be in o1.dot.com(which is a alias)?
AFAIK the installation loads the octopus conf that seems to include "wrong" domain and therefore maybe I could change it manually and try running the upgrade once more?

Project:Barracuda» Octopus
Component:Aegir Master Instance» Aegir Hostmaster

As I have requested already, please change _DEBUG_MODE=NO to _DEBUG_MODE=YES and then run the upgrade and *attach* (don't post inline!) the full output.

Also, please don't *edit* the issue or comments, because you are introducing a total chaos and confusion. All updates should be posted as *new* comments *only*.

StatusFileSize
new5.02 KB

Ok, sorry :)

I did try the debug mode and posted the "results" in the previous comment.
In short:
Upgrade got hanging and nothing happened.
Just a notice "waiting 5 sec" and hang(waited for 30min).
Went to see logfile which ended in: Do you want to proceed? (which is the first step in upgrade)

If I now try the upgrade again it will say that it is already up to date.

However, main point is the most recent "inline" log that you can find from attachments below :) It declares that there was errors in the first place.

OK, could you please try the same with _DEBUG_MODE=YES? But there is something else, which looks much worse:

/var/backups/OCTOPUS.sh.o1: line 589: read: read error: 0: Input/output error

I have never seen anything like that before. Something seems to be seriously broken or untypical on this system if such a simple and basic thing doesn't work.

What is the parent system there and what are its default locales? I mean the result of locale command.

StatusFileSize
new131.48 KB

OK, I tried following:
1. _DEBUG_MODE=YES
2. wget...
3. bash...
4.octopus up-stable all both log
5.OCTOPUS upgrade completed Bye
6. log file as attachment

Locale is fi_FI.utf-8.

This log shows successful auto-recovery from previous failed upgrade:

Octopus [pe 19.4.2013 22.16.09 +0300] ==> UPGRADE B: Testing previous install...
Octopus [pe 19.4.2013 22.16.09 +0300] ==> UPGRADE B: Testing previous install... OPS, zombie found, moving it to backups...
Octopus [pe 19.4.2013 22.16.11 +0300] ==> UPGRADE B: Testing previous install again after removing zombie...
Octopus [pe 19.4.2013 22.16.13 +0300] ==> UPGRADE B: Hostmaster STATUS: upgrade start
Octopus [pe 19.4.2013 22.16.15 +0300] ==> UPGRADE B: Running hostmaster-migrate, please wait...

This time it worked just fine:

Octopus [pe 19.4.2013 22.16.50 +0300] ==> UPGRADE B: Hostmaster STATUS: upgrade completed
Octopus [pe 19.4.2013 22.16.50 +0300] ==> UPGRADE B: Simple check if Aegir upgrade is successful
Octopus [pe 19.4.2013 22.16.52 +0300] ==> UPGRADE B: Aegir upgrade test result: OK

That's what it seems like.

At the moment
front end -> 404
site-> 500

Nginx error log declares "permission denied" when trying to access front end:
"/var/aegir/host_master/003/" (13: permission denied)
"/var/aegir/host_master/003/index.php" (13: permission denied)
"/var/aegir/host_master/003/sites/.../files/..." (13: permission denied)
...

If I try to change the permissions into e.g. rwx for aegir user and group in var/aegir/host_master/003/... it won't no longer give errors related to sites/.../files/.
For the index.php/front end part it turns into a error 500 but nginx doesn't show any erros in the log.
Obviously permission denied was solved and now something else gets in the way.

Could you please run another barracuda up-stable upgrade after enabling drush debugging in /root/.barracuda.cnf and attach the output? I guess that it needs to fix the master instance as it fixed the satellite instance.

Version:» 6.x-2.0-rc8
Status:Postponed (maintainer needs more info)» Closed (cannot reproduce)

Closing due to a lack of response for a long time.

Issue summary:View changes

edits and updates