New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Unexpected behavior with @aliases run from inside a local drupal docroot #313

Closed
joshkoenig opened this Issue Dec 7, 2013 · 24 comments

Comments

Projects
None yet
6 participants
@joshkoenig

joshkoenig commented Dec 7, 2013

Not sure if this is by design or not, but I get unexpected behavior when running drush at an alias if I do it from within a drush installation. It definitely connects to the remote instance, but then seems to attempt to use the local install to fulfill some of the request.

For instance, status shows very little info, and watchdog-show fails due to inability to bootstrap. Once I move out of the drupal docroot locally, it's back to normal.

@weitzman

This comment has been minimized.

Show comment
Hide comment
@weitzman

weitzman Dec 8, 2013

Member

Only a few commands do local processing despite the presence of an alias (e.g. core-ssh, site-set, ...).

Could you post the command you are running along with its output when run with --debug. If the command is not in Drush core (what release are you using?), post the command definition as well. Also, please post your alias definition. Thanks.

Member

weitzman commented Dec 8, 2013

Only a few commands do local processing despite the presence of an alias (e.g. core-ssh, site-set, ...).

Could you post the command you are running along with its output when run with --debug. If the command is not in Drush core (what release are you using?), post the command definition as well. Also, please post your alias definition. Thanks.

@joshkoenig

This comment has been minimized.

Show comment
Hide comment
@joshkoenig

joshkoenig Dec 9, 2013

The command I was using was "status":

drush -dv --strict=0 @pantheon.outlandish-josh.live status
Bootstrap to phase 0. [0 sec, 1.88 MB]                               [bootstrap]
Drush bootstrap phase : _drush_bootstrap_drush() [0.01 sec, 1.99 MB] [bootstrap]
Loaded alias @pantheon.outlandish-josh.live from file /Users/joshk/.drush/pantheon.aliases.drushrc.php [0.01    [notice]
sec, 2 MB]
Undefined index: root sitealias.inc:1168 [0.01 sec, 2.01 MB]                                                    [notice]
Cache HIT cid: 6.1.0-commandfiles-0-b09f40227cd1739fc6b4e461ccfa7663 [0.01 sec, 2.02 MB]                         [debug]
Begin redispatch via invoke process [0.11 sec, 6.45 MB]                                                         [notice]
Backend invoke: ssh -p 2222 -o "AddressFamily inet"                                                            [command]
live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20@appserver.live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20.drush.in 'drush
 --strict=0 --verbose --debug --uri=outlandishjosh.com
--db-url='\''mysql://pantheon:119436e2248847fb99e7e3f422a06806@dbserver.live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20.drush.in:10000/pantheon'\''
--root=/Users/joshk/sandbox/outlandish-josh  core-status 2>&1' 2>&1 [0.2 sec, 6.49 MB]
executing ssh -p 2222 -o "AddressFamily inet"                                                                   [notice]
live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20@appserver.live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20.drush.in 'drush
 --strict=0 --verbose --debug --uri=outlandishjosh.com
--db-url='\''mysql://pantheon:119436e2248847fb99e7e3f422a06806@dbserver.live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20.drush.in:10000/pantheon'\''
--root=/Users/joshk/sandbox/outlandish-josh  core-status 2>&1' 2>&1 [0.2 sec, 6.49 MB]
Calling proc_open(ssh -p 2222 -o "AddressFamily inet" live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20@appserver.live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20.drush.in 'drush  --strict=0 --verbose --debug --uri=outlandishjosh.com --db-url='\''mysql://pantheon:119436e2248847fb99e7e3f422a06806@dbserver.live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20.drush.in:10000/pantheon'\'' --root=/Users/joshk/sandbox/outlandish-josh  core-status 2>&1' 2>&1);
Bootstrap to phase 0. [0.01 sec, 3.54 MB]                            [bootstrap]
Drush bootstrap phase : _drush_bootstrap_drush() [0.01 sec, 3.74 MB] [bootstrap]
Loading drushrc                                                      [bootstrap]
"/srv/bindings/af12454fa2374c6e96a9da839815d649/drushrc.php" into
"custom" scope. [0.02 sec, 3.74 MB]
Loading drushrc                                                      [bootstrap]
"/srv/bindings/af12454fa2374c6e96a9da839815d649/drushrc.php" into
"custom" scope. [0.02 sec, 3.74 MB]
Cache HIT cid: 5.10.0-commandfiles-0-3b6dd35b4aa92a55e325010e322287d2    [debug]
[0.02 sec, 3.75 MB]
Bootstrap to phase 0. [0.05 sec, 8.51 MB]                            [bootstrap]
Found command: core-status (commandfile=core) [0.06 sec, 8.51 MB]    [bootstrap]
 PHP configuration   :  /srv/bindings/af12454fa2374c6e96a9da839815d649/php53.in
                        i
                        /srv/bindings/af12454fa2374c6e96a9da839815d649/php53.in
                        i
 Drush version       :  5.10.0
 Drush               :  /srv/bindings/af12454fa2374c6e96a9da839815d649/drushrc.
 configuration          php

Command dispatch complete [0.08 sec, 8.52 MB]                           [notice]
Peak memory usage was 10.04 MB [0.08 sec, 8.52 MB]                      [memory]
Invoke process is complete [1.41 sec, 6.48 MB]                                                                  [notice]

Here's my alias:

$aliases['outlandish-josh.live'] = array(
  'uri' => 'outlandishjosh.com',
  'db-url' => 'mysql://pantheon:BLANK@dbserver.live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20.drush.in:10000/pantheon',
  'db-allows-remote' => TRUE,
  'remote-host' => 'appserver.live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20.drush.in',
  'remote-user' => 'live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20',
  'ssh-options' => '-p 2222 -o "AddressFamily inet"',
  'path-aliases' => array(
    '%files' => 'code/sites/default/files',
  ),
);

joshkoenig commented Dec 9, 2013

The command I was using was "status":

drush -dv --strict=0 @pantheon.outlandish-josh.live status
Bootstrap to phase 0. [0 sec, 1.88 MB]                               [bootstrap]
Drush bootstrap phase : _drush_bootstrap_drush() [0.01 sec, 1.99 MB] [bootstrap]
Loaded alias @pantheon.outlandish-josh.live from file /Users/joshk/.drush/pantheon.aliases.drushrc.php [0.01    [notice]
sec, 2 MB]
Undefined index: root sitealias.inc:1168 [0.01 sec, 2.01 MB]                                                    [notice]
Cache HIT cid: 6.1.0-commandfiles-0-b09f40227cd1739fc6b4e461ccfa7663 [0.01 sec, 2.02 MB]                         [debug]
Begin redispatch via invoke process [0.11 sec, 6.45 MB]                                                         [notice]
Backend invoke: ssh -p 2222 -o "AddressFamily inet"                                                            [command]
live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20@appserver.live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20.drush.in 'drush
 --strict=0 --verbose --debug --uri=outlandishjosh.com
--db-url='\''mysql://pantheon:119436e2248847fb99e7e3f422a06806@dbserver.live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20.drush.in:10000/pantheon'\''
--root=/Users/joshk/sandbox/outlandish-josh  core-status 2>&1' 2>&1 [0.2 sec, 6.49 MB]
executing ssh -p 2222 -o "AddressFamily inet"                                                                   [notice]
live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20@appserver.live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20.drush.in 'drush
 --strict=0 --verbose --debug --uri=outlandishjosh.com
--db-url='\''mysql://pantheon:119436e2248847fb99e7e3f422a06806@dbserver.live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20.drush.in:10000/pantheon'\''
--root=/Users/joshk/sandbox/outlandish-josh  core-status 2>&1' 2>&1 [0.2 sec, 6.49 MB]
Calling proc_open(ssh -p 2222 -o "AddressFamily inet" live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20@appserver.live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20.drush.in 'drush  --strict=0 --verbose --debug --uri=outlandishjosh.com --db-url='\''mysql://pantheon:119436e2248847fb99e7e3f422a06806@dbserver.live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20.drush.in:10000/pantheon'\'' --root=/Users/joshk/sandbox/outlandish-josh  core-status 2>&1' 2>&1);
Bootstrap to phase 0. [0.01 sec, 3.54 MB]                            [bootstrap]
Drush bootstrap phase : _drush_bootstrap_drush() [0.01 sec, 3.74 MB] [bootstrap]
Loading drushrc                                                      [bootstrap]
"/srv/bindings/af12454fa2374c6e96a9da839815d649/drushrc.php" into
"custom" scope. [0.02 sec, 3.74 MB]
Loading drushrc                                                      [bootstrap]
"/srv/bindings/af12454fa2374c6e96a9da839815d649/drushrc.php" into
"custom" scope. [0.02 sec, 3.74 MB]
Cache HIT cid: 5.10.0-commandfiles-0-3b6dd35b4aa92a55e325010e322287d2    [debug]
[0.02 sec, 3.75 MB]
Bootstrap to phase 0. [0.05 sec, 8.51 MB]                            [bootstrap]
Found command: core-status (commandfile=core) [0.06 sec, 8.51 MB]    [bootstrap]
 PHP configuration   :  /srv/bindings/af12454fa2374c6e96a9da839815d649/php53.in
                        i
                        /srv/bindings/af12454fa2374c6e96a9da839815d649/php53.in
                        i
 Drush version       :  5.10.0
 Drush               :  /srv/bindings/af12454fa2374c6e96a9da839815d649/drushrc.
 configuration          php

Command dispatch complete [0.08 sec, 8.52 MB]                           [notice]
Peak memory usage was 10.04 MB [0.08 sec, 8.52 MB]                      [memory]
Invoke process is complete [1.41 sec, 6.48 MB]                                                                  [notice]

Here's my alias:

$aliases['outlandish-josh.live'] = array(
  'uri' => 'outlandishjosh.com',
  'db-url' => 'mysql://pantheon:BLANK@dbserver.live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20.drush.in:10000/pantheon',
  'db-allows-remote' => TRUE,
  'remote-host' => 'appserver.live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20.drush.in',
  'remote-user' => 'live.3ecd4d40-bdf2-4e52-a519-7e697ecdfe20',
  'ssh-options' => '-p 2222 -o "AddressFamily inet"',
  'path-aliases' => array(
    '%files' => 'code/sites/default/files',
  ),
);
@greg-1-anderson

This comment has been minimized.

Show comment
Hide comment
@greg-1-anderson

greg-1-anderson Dec 10, 2013

Member

Why is there no 'root' defined in the alias file above?

Member

greg-1-anderson commented Dec 10, 2013

Why is there no 'root' defined in the alias file above?

@greg-1-anderson

This comment has been minimized.

Show comment
Hide comment
@greg-1-anderson

greg-1-anderson Dec 10, 2013

Member

I haven't tried to reproduce this yet, but I think that what is going on here is that Drush requires a Drupal root; if there isn't one in the alias, then Drush is going to try to look one up from the cwd. This is probably what's causing your problem.

If you have a legitimate reason for not putting the Drupal root into your site alias, then we can probably fix Drush up to allow this use case. It's not something we've considered prior to now, though, and if it works, it's more be coincidence than design.

Is the Drupal root in this instance taken to be the home directory of the user specified by 'remote-user'?

Member

greg-1-anderson commented Dec 10, 2013

I haven't tried to reproduce this yet, but I think that what is going on here is that Drush requires a Drupal root; if there isn't one in the alias, then Drush is going to try to look one up from the cwd. This is probably what's causing your problem.

If you have a legitimate reason for not putting the Drupal root into your site alias, then we can probably fix Drush up to allow this use case. It's not something we've considered prior to now, though, and if it works, it's more be coincidence than design.

Is the Drupal root in this instance taken to be the home directory of the user specified by 'remote-user'?

@joshkoenig

This comment has been minimized.

Show comment
Hide comment
@joshkoenig

joshkoenig Dec 10, 2013

Yeah, the remote root absolute path is subject to change without notice, but it will always be '~/code' or '$HOME/code'.

I am thinking on some ways to have a reliable root path for other reasons as well though.

joshkoenig commented Dec 10, 2013

Yeah, the remote root absolute path is subject to change without notice, but it will always be '~/code' or '$HOME/code'.

I am thinking on some ways to have a reliable root path for other reasons as well though.

@greg-1-anderson

This comment has been minimized.

Show comment
Hide comment
@greg-1-anderson

greg-1-anderson Dec 10, 2013

Member

Okay, so how does the remote instance of Drush get the Drupal root for the target site? Is it set in $options['root'] of the drushrc.php file in the user's home directory, or something like that?

Member

greg-1-anderson commented Dec 10, 2013

Okay, so how does the remote instance of Drush get the Drupal root for the target site? Is it set in $options['root'] of the drushrc.php file in the user's home directory, or something like that?

@joshkoenig

This comment has been minimized.

Show comment
Hide comment
@joshkoenig

joshkoenig Dec 10, 2013

It is "just working" because we drop the user into the docroot on the remote side.

joshkoenig commented Dec 10, 2013

It is "just working" because we drop the user into the docroot on the remote side.

greg-1-anderson pushed a commit to greg-1-anderson/drush that referenced this issue Dec 11, 2013

Greg Anderson
#313: Don't create a @self alias before the first bootstrap has compl…
…eted; otherwise, this could cause settings from a Drupal site at the cwd to contaminate the settings in a remote site alias if it has no 'root' specified.
@greg-1-anderson

This comment has been minimized.

Show comment
Hide comment
@greg-1-anderson

greg-1-anderson Dec 11, 2013

Member

Try the commit above; it should clear things up for you. I haven't tested this thoroughly, so there may be side effects; I think it's pretty likely to work, though.

Member

greg-1-anderson commented Dec 11, 2013

Try the commit above; it should clear things up for you. I haven't tested this thoroughly, so there may be side effects; I think it's pretty likely to work, though.

@joshkoenig

This comment has been minimized.

Show comment
Hide comment
@joshkoenig

joshkoenig Dec 11, 2013

Well I don't want you to commit anything without testing, but it certainly fixes my particular problem.

Many thanks!!!

joshkoenig commented Dec 11, 2013

Well I don't want you to commit anything without testing, but it certainly fixes my particular problem.

Many thanks!!!

@weitzman

This comment has been minimized.

Show comment
Hide comment
@weitzman

weitzman Dec 11, 2013

Member

Code looks fine. I'm pretty unsure what the repercussions are as well. Greg can commit this when he is satisfied with it. No idea if further test coverage is warranted. Could be hard to test.

Member

weitzman commented Dec 11, 2013

Code looks fine. I'm pretty unsure what the repercussions are as well. Greg can commit this when he is satisfied with it. No idea if further test coverage is warranted. Could be hard to test.

greg-1-anderson added a commit to greg-1-anderson/drush that referenced this issue Dec 11, 2013

@greg-1-anderson

This comment has been minimized.

Show comment
Hide comment
@greg-1-anderson

greg-1-anderson Dec 11, 2013

Member

The theory with d5604d1 and 5e3248d is that all of the same code executes prior to the first call to any commandfile callback, so there should be no change in behavior for any existing core or contrib Drush commands, so long as the Drush bootstrap is still working after these changes.

d5604d1 did break the bootstrap, though, because it defined the @self alias too late. 5e3248d fixes this problem by defining @self a little sooner, and now all of the unit tests pass again.

@joshkoenig: Does this still work for Pantheon? It should be equivalent.

I think this is ready to be committed to master now, but the bigger question is whether we should also backport to 6.x and 5.x. As I said, in theory, this change should be invisible to Drush commandfiles, so by that criteria, it is eligible for backporting. As Moshe said, though, it is difficult to test this, beyond the level of insuring that the unit tests we already have still pass. While it is unlikely to cause a problem, it would be quite surprising if there was some problem that we did not find that affected someone doing a minor update to the next stable release of Drush 5.

The more conservative thing to do, then, would be to commit only to master and 6.x, and leave 5.x alone. Does Pantheon work with Drush 6 yet, though?

Member

greg-1-anderson commented Dec 11, 2013

The theory with d5604d1 and 5e3248d is that all of the same code executes prior to the first call to any commandfile callback, so there should be no change in behavior for any existing core or contrib Drush commands, so long as the Drush bootstrap is still working after these changes.

d5604d1 did break the bootstrap, though, because it defined the @self alias too late. 5e3248d fixes this problem by defining @self a little sooner, and now all of the unit tests pass again.

@joshkoenig: Does this still work for Pantheon? It should be equivalent.

I think this is ready to be committed to master now, but the bigger question is whether we should also backport to 6.x and 5.x. As I said, in theory, this change should be invisible to Drush commandfiles, so by that criteria, it is eligible for backporting. As Moshe said, though, it is difficult to test this, beyond the level of insuring that the unit tests we already have still pass. While it is unlikely to cause a problem, it would be quite surprising if there was some problem that we did not find that affected someone doing a minor update to the next stable release of Drush 5.

The more conservative thing to do, then, would be to commit only to master and 6.x, and leave 5.x alone. Does Pantheon work with Drush 6 yet, though?

@joshkoenig

This comment has been minimized.

Show comment
Hide comment
@joshkoenig

joshkoenig Dec 16, 2013

We're working on making 6.x the "main" drush option on the Platform now. Hoping to switch early in the new year. However, this bug manifests on the client side.

I'm actually not able to replicate the problem with the tip of master, even before applying 5e3248d, so it may not be necessary.

The issue is there in 6.2.0 though.

joshkoenig commented Dec 16, 2013

We're working on making 6.x the "main" drush option on the Platform now. Hoping to switch early in the new year. However, this bug manifests on the client side.

I'm actually not able to replicate the problem with the tip of master, even before applying 5e3248d, so it may not be necessary.

The issue is there in 6.2.0 though.

@greg-1-anderson

This comment has been minimized.

Show comment
Hide comment
@greg-1-anderson

greg-1-anderson Dec 16, 2013

Member

Okay, I will commit 5e3248d to master and 6.x per #313 (comment). I guess we'll leave the 5.x branch alone, then?

Member

greg-1-anderson commented Dec 16, 2013

Okay, I will commit 5e3248d to master and 6.x per #313 (comment). I guess we'll leave the 5.x branch alone, then?

@weitzman

This comment has been minimized.

Show comment
Hide comment
@weitzman

weitzman Jun 17, 2014

Member

@greg-1-anderson - Did this code ever get committed to master and 6.x?

Member

weitzman commented Jun 17, 2014

@greg-1-anderson - Did this code ever get committed to master and 6.x?

greg-1-anderson added a commit that referenced this issue Jun 26, 2014

greg-1-anderson added a commit that referenced this issue Jun 26, 2014

@greg-1-anderson

This comment has been minimized.

Show comment
Hide comment
@greg-1-anderson

greg-1-anderson Jun 26, 2014

Member

No, this was missed. I committed 5e3248d to master and 6.x as 3ac904e just now.

Member

greg-1-anderson commented Jun 26, 2014

No, this was missed. I committed 5e3248d to master and 6.x as 3ac904e just now.

@greg-1-anderson

This comment has been minimized.

Show comment
Hide comment
@greg-1-anderson

greg-1-anderson Jul 1, 2014

Member

Oh, darn. This caused a subtle failure in the unit tests because the order of evaluation for DRUPALROOT/sites/default/drushrc.php and DRUPALROOT/drush/drushrc.php changed. The Drupal-specific config file should load before the site-specific config file, in case the later overrides options in the former.

I fixed this in master with 7c3bd62. This commit changed a long-standing behavior of Drush: any time you reference any alias, even in contexts that do not cause the site to bootstrap (e.g. sql-sync), Drush would load the drushrc.php file located next to the site's settings.php file. Arguably, Drush never should have been this liberal; however, fixing this problem would be a change in behavior.

The behavior change is fairly subtle, on par with 3ac904e. So, the question is now, back out 3ac904e from 6.x, or commit 7c3bd62? One or the other is required to fix the broken 6.x unit tests.

Member

greg-1-anderson commented Jul 1, 2014

Oh, darn. This caused a subtle failure in the unit tests because the order of evaluation for DRUPALROOT/sites/default/drushrc.php and DRUPALROOT/drush/drushrc.php changed. The Drupal-specific config file should load before the site-specific config file, in case the later overrides options in the former.

I fixed this in master with 7c3bd62. This commit changed a long-standing behavior of Drush: any time you reference any alias, even in contexts that do not cause the site to bootstrap (e.g. sql-sync), Drush would load the drushrc.php file located next to the site's settings.php file. Arguably, Drush never should have been this liberal; however, fixing this problem would be a change in behavior.

The behavior change is fairly subtle, on par with 3ac904e. So, the question is now, back out 3ac904e from 6.x, or commit 7c3bd62? One or the other is required to fix the broken 6.x unit tests.

@greg-1-anderson

This comment has been minimized.

Show comment
Hide comment
@greg-1-anderson
Member

greg-1-anderson commented Jul 1, 2014

Ping @weitzman

@weitzman

This comment has been minimized.

Show comment
Hide comment
@weitzman

weitzman Jul 3, 2014

Member

I could go either way. I'll defer to @greg-1-anderson.

Member

weitzman commented Jul 3, 2014

I could go either way. I'll defer to @greg-1-anderson.

@greg-1-anderson

This comment has been minimized.

Show comment
Hide comment
@greg-1-anderson

greg-1-anderson Jul 3, 2014

Member

I committed 7c3bd62 on the 6.x branch, as Drush works better like this.

If there is someone in the world who is using a site-local drushrc.php file to change the behavior of sql-sync or drush rsync (which works for local sites only), then this change could be fairly painful, in that the fix is not very discoverable. Odds are good that there's no one doing that, though, and there's always avenues for support requests here & on stack exchange if it does come up.

Member

greg-1-anderson commented Jul 3, 2014

I committed 7c3bd62 on the 6.x branch, as Drush works better like this.

If there is someone in the world who is using a site-local drushrc.php file to change the behavior of sql-sync or drush rsync (which works for local sites only), then this change could be fairly painful, in that the fix is not very discoverable. Odds are good that there's no one doing that, though, and there's always avenues for support requests here & on stack exchange if it does come up.

@liberatr

This comment has been minimized.

Show comment
Hide comment
@liberatr

liberatr Sep 17, 2014

I noticed that this works fine in 6.3 but not 6.4 - also using Pantheon, so they still don't provide %root. Damien says he fixed it in #821 but I couldn't get it to work, so I downgraded to drush 6.3.

It also behaves slightly differently when you are inside a drupal folder or not.

Just wanted to make sure I noted it here since this page appears when you google the phrase "Command not supported as typed."

liberatr commented Sep 17, 2014

I noticed that this works fine in 6.3 but not 6.4 - also using Pantheon, so they still don't provide %root. Damien says he fixed it in #821 but I couldn't get it to work, so I downgraded to drush 6.3.

It also behaves slightly differently when you are inside a drupal folder or not.

Just wanted to make sure I noted it here since this page appears when you google the phrase "Command not supported as typed."

@joshkoenig

This comment has been minimized.

Show comment
Hide comment
@joshkoenig

joshkoenig Sep 18, 2014

"Command not supported as typed."

In the context of Drush 6.4 this is a totally separate issue:

pantheon-systems/terminus#174

joshkoenig commented Sep 18, 2014

"Command not supported as typed."

In the context of Drush 6.4 this is a totally separate issue:

pantheon-systems/terminus#174

@ForestMars

This comment has been minimized.

Show comment
Hide comment
@ForestMars

ForestMars Aug 17, 2015

This page was suggested as background for Pantheon ticket #43242 ("looks more like a drush problem than a Pantheon problem.")

However given that drush features is apparently/completely broken on Pantheon (per support rep) I'd say it's everybody's problem, at least anyone who is using drush to deploy Drupal. (Really, who does that anyway?)

ForestMars commented Aug 17, 2015

This page was suggested as background for Pantheon ticket #43242 ("looks more like a drush problem than a Pantheon problem.")

However given that drush features is apparently/completely broken on Pantheon (per support rep) I'd say it's everybody's problem, at least anyone who is using drush to deploy Drupal. (Really, who does that anyway?)

@RobLoach

This comment has been minimized.

Show comment
Hide comment
@RobLoach

RobLoach Nov 18, 2015

Contributor

For those of you still following along, @greg-1-anderson wrote a small article about using policy.drush.inc:
https://pantheon.io/blog/fix-drush-site-aliases-policy-file

May help, or not. Here's some ice cream! 🍦

Contributor

RobLoach commented Nov 18, 2015

For those of you still following along, @greg-1-anderson wrote a small article about using policy.drush.inc:
https://pantheon.io/blog/fix-drush-site-aliases-policy-file

May help, or not. Here's some ice cream! 🍦

@greg-1-anderson

This comment has been minimized.

Show comment
Hide comment
@greg-1-anderson

greg-1-anderson Nov 18, 2015

Member

Drush 8 is now the default on Pantheon for new sites, but old sites continue to use Drush 5. Upgrading to Drush 8 is recommended.

Member

greg-1-anderson commented Nov 18, 2015

Drush 8 is now the default on Pantheon for new sites, but old sites continue to use Drush 5. Upgrading to Drush 8 is recommended.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment