I've been comparing unRAID to my other NAS boxes (ReadyNAS, Thecus 5200BR) and there are a number of features missing.

One of which is mail delivery to an administrative address.

 

In preparation I started playing with sendmail.

I installed it and found it to have issues with the root filesystem on ramfs.

Somehow it checks the filesystem space and if the space is low it does not allow queuing of mail.

The issue is,  it does not get the right filesystem size for root on ramfs.

In response, I mounted a tmpfs of 5M each on /var/spool/mqueue and /var/spool/clientmqueue.

Not a terrible thing, but perhaps not what we want.

The good thing is that it keeps stagnant mail from filling up root.

I thought it a bit complicated, but workable.

 

I checked the ReadyNAS and it is using exim named as sendmail.

I installed that as a test too, no difficulties with installation and running (that I could find).

 

When I updated the aliass files so root was

 

root:  [email protected]

 

Mail via sendmail call or via cron jobs worked.

 

So at least the groundwork is done, but is it something that the rest of the group thinks should be a standard feature.

I.E. Mailed alerts to an administrative address.

Heck, even my cheesy router does this!

 

 

Furthermore, veering slightly off topic, I store my mail using iMAP on one of my ITX servers.

I've been thinking of installing dovecot on the uinraid server and setting the mail location to one of the data disks.

This is purely for archival reasons.

I.E. I buy something, get a receipt, I move it from my iMAP folder at my ISP to a local IMAP folder on the archival server. and wala... My mail is saved and accessible for as long as I keep it.

You can even program Thunderbird to automatically move messages with filters. Works great for me.

 

I don't necessarily see(or agree upon) unRAID being a viable Mail environment for receipt, but archiving yes.

 

So is this also something useful to the rest of the group?

Email notification would be a great feature and -- as you noted -- is offered on competing solutions.  My servers are very stable and, so far, trouble-free.  However, without having a notification system, I'm always concerned that something might go wrong & I won't find out about it on a timely basis.  Anything that can be done to address this in a future version would be very appreciated!

I would find this feature useful. I don't want to manually check for errors, bad hard drives, or other issues that can come up. I'd much rather let unRAID run on it's own and when a problem arises, be notified of it via email. This way the only times I'll really need to log in via HTTP or telnet is when I want to add/remove a hard drive or do some sort of in depth check. The way unRAID works now I will constantly be viewing the HTTP interface and every so often logging in via telnet to check disk status'.

I would find this feature useful. I don't want to manually check for errors, bad hard drives, or other issues that can come up. I'd much rather let unRAID run on it's own and when a problem arises, be notified of it via email. This way the only times I'll really need to log in via HTTP or telnet is when I want to add/remove a hard drive or do some sort of in depth check. The way unRAID works now I will constantly be viewing the HTTP interface and every so often logging in via telnet to check disk status'.

If all you are looking for is an alert when your array has a problem, you can add the e-mail alert yourself as described in this thread: http://lime-technology.com/forum/index.php?topic=911.0

 

It is not "sendmail" but is way less to install, and can handle and alert you to most error conditions.

 

Joe L.

Thanks Joe... I'll play around with that tonight when I have some free time

That script works great. I found myself needing more.

 

With a real mail transfer agent I can capture output of my cron jobs to my sysadmin account.

I'm also looking into the future for us to run smartmon tolls smartd which will automatically check the status of your drives.

You can schedule testing automatically and have it alert you if there is a physical drive problem.

 

I also have special syslog monitor scripts that capture odd lines and messages and send them via mail.

 

 

I found another program out there called smtpsend which can do what the netcat does without all of the shell semantics.

It's just a simpler SMTP sending program.  This may suffice.

 

 

Another feature about having a real Mail Transfer Agent is that it will queue the mail if the destination is down.

 

 

exim is pretty light weight compared to others.

 

It's one small package, two updates to a conf file (which I do in the install script) and you are ready to go.

I have it almost ready to publish.

 

If you log mail locally via syslog it doesn't even run the daemon. (this isn't too much help to many, but it can at least capture your cron job output to syslog).