June 1, 200917 yr Forking from the prior thread. http://lime-technology.com/forum/index.php?topic=3839.msg33730#msg33730 There were a number of suggestions posted on possible ways to improve noticing if the array has issues. We currently have one basic method. Review your array's http interface periodically and look for a red mark next to your drive. What ways would you like to be notified if your array has issues. What should be in the standard unRAID distribution. In a number of other NAS devices I own (or have owned) Each had an outbound mail notification in addition to the browser interface. I think BubbaQ's browser plugin was a step in a desktop monitoring solution, as was Joe's suggestion of WhoIsIt21 http://lime-technology.com/forum/index.php?topic=3839.msg33838#msg33838 I personally started to work on Nagios for unRAID, but there was not that much interest. It will be interesting to see what idea's people have and to document the events that should be alerted.
June 1, 200917 yr I'm not a huge fan of RSS (mostly cause I do not use it). I think e-mail alerts are a must. The ability to set it up, and tell it everything you want to be email about is necessary. Anything after email notification is a plus as far as I am concerned. I am a huge fan for Growl for OS X and will soon be messing with Growl for windows. I here they are trying to get a linux version up and running. Something like a heads up notification would be great, but not really necessary.
June 1, 200917 yr It all comes down to proper alert classifications. The method of delivery is less important. I operate quite a number of firewall clusters and i easily get 500 emails a day from them alone... which is quite frankly useless. However with simple scaled tagging RSS/EMAIL and everything else becomes much simpler. i.e. INFO - Disk 99% full in the backend with me ticking RSS, beep and email on INFO or perhaps i just want INFO in RSS and ALERT in beep. you get the point, everyone will prefer some differernt means of alerting front end.... the back end should be the clever bit
June 1, 200917 yr It all comes down to proper alert classifications. The method of delivery is less important. I operate quite a number of firewall clusters and i easily get 500 emails a day from them alone... which is quite frankly useless. However with simple scaled tagging RSS/EMAIL and everything else becomes much simpler. i.e. INFO - Disk 99% full in the backend with me ticking RSS, beep and email on INFO or perhaps i just want INFO in RSS and ALERT in beep. you get the point, everyone will prefer some differernt means of alerting front end.... the back end should be the clever bit Agree completely. I just thought of another thing that would be good to have in general. Some type of automatic saving of the syslog when stuff like this happens. Most new users first instinct is to restart the computer and see if that fixes the problem, the problem with that is that we don't get anything to look at to help them diagnose the problem. Just whenever the Restart/Shutdown/Power button is pushed a simple script is run to save a copy of the syslog to the flash drive. That way it is there and we have something to look and go from when help is asked for.
June 2, 200917 yr Audible beeps and email notifications giving the status of: 1) Array status (is the array protected) 2) disk temp 3) Smart monitoring (is a disk ready to fail?) 4) cpu temp/case temp 5) Fan failure In addition, an optional email notification / periodic chirp letting you know that a parity check hasn't been performed in awhile might be helpful.
June 2, 200917 yr Author I just thought of another thing that would be good to have in general. Some type of automatic saving of the syslog when stuff like this happens. Most new users first instinct is to restart the computer and see if that fixes the problem, the problem with that is that we don't get anything to look at to help them diagnose the problem. Just whenever the Restart/Shutdown/Power button is pushed a simple script is run to save a copy of the syslog to the flash drive. That way it is there and we have something to look and go from when help is asked for. I think this is possible with my powerdown package. It automatically saves 10 syslogs on the flash. I had set up a patch to the log rotate that would also call the powerdown savesyslog option. This would save the current syslog in the rotating 10 syslogs on flash just before rotation. People did not seem very interested in it so I did not pursue it. I could review this again if there is interest.
June 2, 200917 yr Audible beeps and email notifications giving the status of: 3) Smart monitoring (is a disk ready to fail?) BubbaQ's Smart History is a very good addon to have as it graphically displays information about the smart history and does a comparison over time. It would be nice to have something like BubbaQ's Smart History in unRAID 5.0
June 2, 200917 yr I think there is a distinction between general notification of array issues (e.g., low on disk space) and notifications of errors (e.g., a disk has been disabled). Perhaps a unified model could address both, but I doubt Tom is going to add any elaborate polling system to unRAID. I think that user developed scripts should step is to provide what I'd call "enhanced notification" for those that want it. But basic notification of disks being disabled need to be in the base product. Perhaps there are other situations that rise to this level. I thnk that beeping, email notification, and a network pushed alert option should exist. The GUI should also have a way of communicating these alerts to the user in a dramatic way, and point users to resources on what they should do. The key to being effective is that these features not be overused. If users are getting emails all of the time from unRAID they'll go unread. If alert boxes are popping up to communicate non-critical info they will be disabled. But if user get notified only once in a while and when they do get them it is important, it will serve its purpose.
June 2, 200917 yr Author Re: RSS. How many people use Thunderbird or a Firefox plugin for RSS? Just curious as I believe with emhttp's ability to read logs in /var/log, an rss file of events could be created of the last 10 events allowing your browser or email reader to poll that way. Re: Growl I'm not that big of a fan. I could see how some people could love it. Seems larger then needed. I probably need to learn more about it. Re: Whoisit. I'm liking this as a notification tool, Not sure how many people would want to install it. Works well, simple network interface. What's missing??? Is there source code available. Will there come a time when it is not supported? So far many of the mentions are client notifications. I think consistent reporting, and monitoring via API is needed first. If many of the warning tools used syslog and logger with consistent message alert conventions, then this single file can be used by other applications to send out the client alerts. For example, I have a syslog checker on my system that scrapes the syslog, greps out what I need and don't need leaving a condensed set of lines for which I choose or have not prepared for. Also if unRAID used a more advanced syslog daemon like syslog NG messges could be filtered directly to applications.
June 2, 200917 yr I'll upgrade FEMUR with more bells and whistles after the new flavor of unRAID with php/webserver is released.
June 2, 200917 yr I just thought of another thing that would be good to have in general. Some type of automatic saving of the syslog when stuff like this happens. Most new users first instinct is to restart the computer and see if that fixes the problem, the problem with that is that we don't get anything to look at to help them diagnose the problem. Just whenever the Restart/Shutdown/Power button is pushed a simple script is run to save a copy of the syslog to the flash drive. That way it is there and we have something to look and go from when help is asked for. I think this is possible with my powerdown package. It automatically saves 10 syslogs on the flash. I had set up a patch to the log rotate that would also call the powerdown savesyslog option. This would save the current syslog in the rotating 10 syslogs on flash just before rotation. People did not seem very interested in it so I did not pursue it. I could review this again if there is interest. I have a powerdown package installed (honestly not sure which version). All I know is that if I do a control+alt+delete I get a clean powerdown. I think that process might also save a copy of the syslog to the flash. I guess what I am saying is that this syslog saving before powerdown, and saving before log rotate, should become the standard. There are a lot of scripts on this board that have been developed and worked on by many of the people smarter then me. Some if not all of them should become part of the standard unRAID install and be configurable from the web interface. A litst of the top of my head: Third party Boot Flash Plugin Architecture - would be great to see this become part of the standard install of unRAID, that way things can become unified on where to install addon's Powerdown package with control+alt+delete and saving of syslogs - can de configurable to save more then 10 from web interface control of blockdev from web interface - not so important but might be nice to have cache_dir - the wonderful script (if you can call it that) by Joe L that keeps in cache the directory listing. This has significantly improved my browsing via Finder on my Mac preclear - Joe L's script that reads and clears disk before they are added to the array. With the growing size of drives it is taking longer and longer to get a drive added to the array some wort of email system - unraid_notify does pretty much want I need it to by e-mailing me every hour and telling me what my server is doing and if there is a problem unmenu - most of the plugins that come with unmenu would be great to have in the stock web interface I understand Tom's want/need to keep is simple. Just my 2 cents
June 13, 200917 yr I just thought of another thing that would be good to have in general. Some type of automatic saving of the syslog when stuff like this happens. Most new users first instinct is to restart the computer and see if that fixes the problem, the problem with that is that we don't get anything to look at to help them diagnose the problem. Just whenever the Restart/Shutdown/Power button is pushed a simple script is run to save a copy of the syslog to the flash drive. That way it is there and we have something to look and go from when help is asked for. I think this is possible with my powerdown package. It automatically saves 10 syslogs on the flash. I had set up a patch to the log rotate that would also call the powerdown savesyslog option. This would save the current syslog in the rotating 10 syslogs on flash just before rotation. People did not seem very interested in it so I did not pursue it. I could review this again if there is interest. I have a powerdown package installed (honestly not sure which version). All I know is that if I do a control+alt+delete I get a clean powerdown. I think that process might also save a copy of the syslog to the flash. I guess what I am saying is that this syslog saving before powerdown, and saving before log rotate, should become the standard. There are a lot of scripts on this board that have been developed and worked on by many of the people smarter then me. Some if not all of them should become part of the standard unRAID install and be configurable from the web interface. A litst of the top of my head: Third party Boot Flash Plugin Architecture - would be great to see this become part of the standard install of unRAID, that way things can become unified on where to install addon's Powerdown package with control+alt+delete and saving of syslogs - can de configurable to save more then 10 from web interface control of blockdev from web interface - not so important but might be nice to have cache_dir - the wonderful script (if you can call it that) by Joe L that keeps in cache the directory listing. This has significantly improved my browsing via Finder on my Mac preclear - Joe L's script that reads and clears disk before they are added to the array. With the growing size of drives it is taking longer and longer to get a drive added to the array some wort of email system - unraid_notify does pretty much want I need it to by e-mailing me every hour and telling me what my server is doing and if there is a problem unmenu - most of the plugins that come with unmenu would be great to have in the stock web interface I understand Tom's want/need to keep is simple. Just my 2 cents i agree that many of those things like unmenu, preclear, dircache, and powerdown should be standard, and easily configurable by the end user via the webinterface. also a long term improvement would be to polish the interface a bit, maybe using ajax or something. (i love the web interface for the tomato firmware on my linux router (linksys wrt54gl). more info on the interface, transfer speeds, connection speeds, everything. more easy info the better
Archived
This topic is now archived and is closed to further replies.