Josh,

It looks like you have active Maintenance and Support on your SmarterMail license, so you may want to submit a support ticket along with an RSAA so that one of our technicians can hop on your server and take a closer look.

If you opt to submit a ticket, you can follow the instructions here:

https://portal.smartertools.com/kb/a1718/remote-server-access-agreements-rsaa.aspx 

Josh, typically when a user's settings.json file becomes corrupted like this it is a sign that the server suffered a power failure or other sudden restart. SmarterMail keeps a variety of JSON configuration files open on a regular basis, and so a sudden shutdown can result in JSON corruption like you're seeing. If no power events were experienced you may want to run file system checks on the domains drive to ensure no disk or controller issues are being experienced. Glad to hear you were able to get this restored and operational again though! Have a great day!

It is possible that this server did experience an unexpected restart at some point, so that is a good explanation.

SUGGESTION: It would be great if SmarterMail would generate an alert anytime it encountered anything unexpected when reading in files. It should be easy for the SmarterMail parser to recognize things like this and alert on them, and I'd much rather be proactive on things like this. 

In the meantime, is there anything in Smartermail equivalent to "scandisk" that I could use to ask SmarterMail to open and check all files looking for latent problems?

Thanks!

-josh

Thanks Josh, thanks for the suggestion regarding notifications on these types of issues, I think this is a great idea. Currently you can navigate to your server's convert-status page (example below) to view any issues in loading the JSON configuration files at the domain or user level: 

Once logged in you will be presented with a domain list along with columns indicating whether an error or warning was encountered in loading that domain. If you click on the domain itself it should then show you the specific account having trouble. I'll go ahead and submit a feature request to have this information more readily available to the administrator interface. 

Lastly, unfortunately there isn't an easy way to scan the entire installation for these issues, but I think this is a great idea as well! I will be discussing both of these ideas with development further this week and I will follow up with you if I learn new information. Thanks again Josh. Have a great rest of your week!

I just hit the exact same issue with a user, and sure enough the json file is just full of null chars. The server did have a power outage (so much for the dc's ups etc grrr). 

Been there, done that - power outage, settings.json files filled with NULL NULL NULL. We tried restoring the settings.json files from backup and that appeared to fix the users, but all kinds of issues and errors occurred while using the WebMail interface for the affected users, until they became unusable. This only occurred with build 7153.

We reinstalled build 7125 and restored from backup all domain files from before the power outage - problem fixed and all working.

Build 7153 is seriously flawed.

Hello Vincent, power failures will cause JSON file corruption due to SmarterMail having these configuration files open when the power failure occurred. If you see corruption in the JSON file you will need to restore them from a backup. For future reference, we keep an Archived Data folder within each user account that should have backups of all pertinent JSON files for each user going back over the last few days, which will allow you to quickly swap in the good copy and then start the services back up again.

Ionel, XML/JSON corruption during power outages has been a known behavior going back several major versions, and this is not new behavior to build 7153. As mentioned above, we now have an Archived Data folder in each user account that helps mitigate these types of issues.

This seems like a design flaw to me.. how is the file being filled with nulls when the power fails? At the very least, if this is a problem then make a copy (in the same folder) when opening the file, and the next time the file is opened, check if it is empty and restore the backup file automatically if corruption is detected. 

Right now we have no way to know if this is happening and on how many accounts, the convert-status page didn't show anything useful (said all was ok). 

This issue caused us a lot of pain, since we had our sales account (which is linked to our CRM) corrupted and missed a bunch of mail for a few weeks before it was discovered. We spent most of yesterday apologising to customers for our tardy response.

We're having this same issue after a hard reset.

Has there been any movement on Josh's suggestion?

SUGGESTION: It would be great if SmarterMail would generate an alert anytime it encountered anything unexpected when reading in files. It should be easy for the SmarterMail parser to recognize things like this and alert on them, and I'd much rather be proactive on things like this.  

I need a way of knowing what accounts are affected.  So far all the users were wiped from one domain and the settings for a user in another domain.  But I'm only learning about these issues from my customers. I want to fix the problems BEFORE the customer notifies me.

I can only agree. In case of a server crash/unclean shutdown, some json files for users, domains settings/folders/etc can be corrupted. 

They can be either empty, filled with garbage, or truncated.

As sysadmins we badly need an integrity check tool for such cases.