Issues Setting Up Advanced Recovery

(Tom Ray) #1

Having some issues getting Advanced Recovery going. Got two servers at two different locations but they are directly on the Internet using the system firewall. Did the SSH key generation between the servers and verified that I can SSH using the keys and login without a password.

When I try to run the quick configuration wizard all it does it tells me there is an SSH issue. The wiki doesn’t have a lot of FAQ on possible errors/steps. I’m curious if anyone else has had this issue and if there was a work around for it? Or perhaps there is a step not outlined in the wiki that needs to be done somewhere else (like Filestore, etc) though I figured a wizard would be dealing with those things.

I’ve also disabled the firewalls on each server and had the same issue. Each time following the “verify SSH” steps in the wiki SSH is verified at the system level but the module continues to say “SSH issue”

(Jared Busch) #2

I’ve not tried to use Advanced Recovery yet, but did you let FreePBX create the SSH key?

Because I tried to use a non-RSA key for a simple SSH backup job and found that FreePBX cannot handle anything more than a basic SA private key.

(Tom Ray) #3

Well I’ve done some more testing and this is what I’ve come up with.

First, I created a Warm Spare backup on the primary server, did all the needed steps. Tried to do a backup, it failed. Couldn’t find /var/spool/asterisk/backup on the secondary server. I looked and sure enough:

drwxrwxr-x. 16 asterisk asterisk  4096 Feb 15 00:06 .
drwxr-xr-x. 11 root     root       167 Feb  9 05:11 ..
-rw-r--r--   1 asterisk asterisk 88250 Feb 18 16:09 173c1676cc42b7e0aeca9338579b3c37.spc
-rw-r--r--   1 asterisk asterisk 71110 Feb 18 16:09 7aa41d5b4c89a3541fcc62095d0c2f61.spc
drwxrwxr-x.  3 asterisk asterisk    24 Feb  9 05:15 cache
drwxrwxr-x.  2 asterisk asterisk    10 Nov 25 14:43 callaccounting
drwxrwxr-x.  2 asterisk asterisk    68 Feb  9 05:15 cxpanel
drwxrwxr-x.  2 asterisk asterisk    10 Nov 25 14:41 fax
drwxrwxr-x.  2 asterisk asterisk    10 Feb 18 16:45 incron
drwxrwxr-x.  2 asterisk asterisk    10 Sep 10 18:03 meetme
drwxrwxr-x.  2 asterisk asterisk    10 Nov 25 14:41 monitor
drwxrwxr-x.  3 asterisk asterisk    26 Feb  9 05:15 oc
drwxrwxr-x.  2 asterisk asterisk    10 Feb  9 05:16 outgoing
drwxr-xr-x.  5 asterisk asterisk    64 Feb  9 05:15 pms
drwxrwxr-x.  2 asterisk asterisk  4096 Feb 16 22:53 sysadmin
drwxrwxr-x.  2 asterisk asterisk    10 Sep 10 18:03 system
drwxrwxr-x.  2 asterisk asterisk    10 Sep 10 18:03 tmp
drwxrwxr-x.  2 asterisk asterisk    10 Sep 10 18:03 voicemail

I then went to the secondary server to double check things and by just visiting the Backup/Restore page (and doing nothing else) resulted in:

drwxrwxr-x. 17 asterisk asterisk  4096 Feb 18 16:50 .
drwxr-xr-x. 11 root     root       167 Feb  9 05:11 ..
-rw-r--r--   1 asterisk asterisk 88250 Feb 18 16:09 173c1676cc42b7e0aeca9338579b3c37.spc
-rw-r--r--   1 asterisk asterisk 71110 Feb 18 16:09 7aa41d5b4c89a3541fcc62095d0c2f61.spc
**drwxr-xr-x   2 asterisk asterisk    10 Feb 18 16:50 backup** << Notice that.
drwxrwxr-x.  3 asterisk asterisk    24 Feb  9 05:15 cache
drwxrwxr-x.  2 asterisk asterisk    10 Nov 25 14:43 callaccounting
drwxrwxr-x.  2 asterisk asterisk    68 Feb  9 05:15 cxpanel
drwxrwxr-x.  2 asterisk asterisk    10 Nov 25 14:41 fax
drwxrwxr-x.  2 asterisk asterisk    10 Feb 18 16:45 incron
drwxrwxr-x.  2 asterisk asterisk    10 Sep 10 18:03 meetme
drwxrwxr-x.  2 asterisk asterisk    10 Nov 25 14:41 monitor
drwxrwxr-x.  3 asterisk asterisk    26 Feb  9 05:15 oc
drwxrwxr-x.  2 asterisk asterisk    10 Feb  9 05:16 outgoing
drwxr-xr-x.  5 asterisk asterisk    64 Feb  9 05:15 pms
drwxrwxr-x.  2 asterisk asterisk  4096 Feb 16 22:53 sysadmin
drwxrwxr-x.  2 asterisk asterisk    10 Sep 10 18:03 system
drwxrwxr-x.  2 asterisk asterisk    10 Sep 10 18:03 tmp
drwxrwxr-x.  2 asterisk asterisk    10 Sep 10 18:03 voicemail

OK cool, the directory exists. So now I try to do the backup again. I get another error. While this time the backup was sent to the secondary server the restore process failed to work. So I could to validate things:

-rw-r--r-- 1 root root 717387 Feb 18 16:53 /var/spool/asterisk/backup/20210218-115305-1613667185-15.0.17.17-193674846.tar.gz

Backup exists. So then I changed the restore method from SSH to API, put in all the API details. Run a new backup, the file is transferred by SSH and the API triggers the restore process, it works fine.

So basically, the primary can log into the secondary via SSH keys from both the CLI and via the Warm Spare Backup process. I still am getting SSH errors from Advanced Recovery so I’m starting to point the issue at the module itself since everything else that requires the SSH key based login between two machines works just fine.

(Lorne Gaetz) #4

If you haven’t already done so, can you open a Support ticket please Tom.
https://wiki.sangoma.com/display/FPAS/How+To+Open+A+Support+Ticket

(Tom Ray) #5

Well this is a 30 day trial of Advanced Recovery for another party, I’m just doing the setup. The admin has spoken with their sales rep about the issue and asked for some assistance, the response was that support credits needed to be purchased. They weren’t thrilled with that response which lead to my warm spare testing.

So there is no included support right now. If you still think opening a ticket is going to get something down without support credits, I will tell the admin to open one since all this is under his Sangoma account (deployment IDs, licenses, etc).

(Tom Ray) #6

So just a little update. A ticket was opened on Thursday about this issue, it took until about 40 minutes ago to get a response that it would be assigned to an engineer. This is after two days of me asking for an update.

However, since there were modules updated in the last few days including another update to Core yesterday (.51) I went into the primary server and updated all the modules that needed to be updated. Now it is important to note, I only did this on the primary server not the secondary server. After the update was completed I tested Advanced Recovery and BOOM! SSH connection is made to the server I kept getting errors that SSH wasn’t setup right. The recovery builds out the primary and the secondary with no issue.

At this point I delete everything in Advanced Recovery on the primary server, this basically resets it to run the configuration wizard again. I jump over the secondary and try to connect back to the primary under Advanced Recovery and bam, “The server does not have SSH access configured properly” so it can’t connect to the system.

So what did I just do, I updated the secondary server and tested again. Lo and behold, I can connect to the primary server via Advanced Recovery so clearly these errors about the other server having SSH misconfigured were false as it only worked when the system issuing the SSH connection via Advanced Recovery was updated to most recent modules released in the last few days.

This was clearly an issue with the modules and something over the releases from the last couple days fixed the issue. Since things like Framework and Core are base modules that almost everything else depends on (perhaps System Admin too) I have to ask, when these modules are updated the QA process does require that all the other modules that depend on them are tested too? Because this seems to be a theme where one module is updated and other things that depend on it aren’t tested and those break.

(Tom Ray) #7

Oh and BTW, who thought this was a great idea for layout on the wizard?

AdvRecoverySetup

You really could have widened the window so things don’t look all jammed in and out of alignment. Oh and of course then there’s this:

AdvRecoverySetup01

Notice how that section becomes scrollable because the tooltip display content? Yeah, so when you hover away from the tooltip to scroll down to read the full tooltip the content becomes hidden again. So good job on that, way to make it helpful for users.

@ncorbic You guys need to be better at this. Period.