I just started off with unRAID and my, there's a lot of stuff to read! So much that sometimes I'm confused where to go next! So I have a list of questions that came up while I was reading to hopefully clear up some of the confusing I have.

 

I have unRAID 4.7 with 3 data drives and will add the parity drive once it is ready.

 

1) I read about the clean powerdown script. What if I want to power down cleanly via the web interface, is there something I need to install?

 

2) I know pressing the 'restore' button is a big no-no. What is the correct procedure in recovering data when a drive fails, and what are some good ways of monitoring drives so we know before they are about to fail? Are there any scripts that will do this and preferably send email notifications?

 

2.5) I've installed monthly parity checks via cron. Are there any other maintenance-related procedures I should be aware of?

 

3) I installed unRAID-Web via unMENU but can't seem to access it via tower:89; I get 400 Bad Request, but I CAN access the web server at tower:2197; is there something I have to configure?

 

4) I recently learned that I should never use the "stop array" button in unMENU and always use the one in unRAID main menu instead. Are there any other similar "don't"'s that a newbie should know about?

 

5) I am running 4.7; I've been reading some legacy threads about increasing the read lookahead to increase read speeds, is this still necessary in current versions?

 

6) I went through the list of packages in unMENU and a lot of them are very interesting and "nice to have", however would installing a ton of these packages onto cron jobs make the server noticeably slower?

 

7) I read through the thread about 'go' scripts and was somewhat confused. http://lime-technology.com/forum/index.php?topic=5976.0 In a nutshell, what are some tweaks that are must have or highly recommended?

 

Some people here love the Supermicro cards, whereas others opt for the ones from Monoprice. What's the difference, performance wise?

 

9) I'm looking to automate, say, weekly or monthly backup of a small subset of important data to another location, either a computer, via web, or even Dropbox. What is possible with unRAID and what is the most practical solution?

 

10) Is there anything similar to jDownloader that is supported by unRAID?

 

Lastly, I would also appreciate if there are anything you feel are "must-reads" for a newbie and wished someone told you when you first started using unRAID.

 

Thanks

I just started off with unRAID and my, there's a lot of stuff to read! So much that sometimes I'm confused where to go next! So I have a list of questions that came up while I was reading to hopefully clear up some of the confusing I have.

 

I have unRAID 4.7 with 3 data drives and will add the parity drive once it is ready.

 

1) I read about the clean powerdown script. What if I want to power down cleanly via the web interface, is there something I need to install?

No, not unless you install a user add-on that keeps a disk busy.  If a disk is busy, the "Stop" button on the web interface will wait (almost forever) for the disk to become "not busy" so it can be un-mounted and the array cleanly stopped.   As long as no disks are "busy" you can stop the array with the button on the web-interface and then power down with the button on the web-interface after the array is "Stopped"

 

2) I know pressing the 'restore' button is a big no-no. What is the correct procedure in recovering data when a drive fails, and what are some good ways of monitoring drives so we know before they are about to fail? Are there any scripts that will do this and preferably send email notifications?

There has been no "restore" button for several releases.  It was replaced with an "initconfig" command in the latter 4.5/6/7 series, and is different again in the latest 5.0beta release.   If a drive fails, step 1, capture a system log.   If you do not know why it failed you are likely to act incorrectly.   It could be as simple as a loose cable, or as complicated as an intermittent piece of hardware.  In NO case will simply fixing a loose connect make the array use the disk again.  The drive was removed from service because a write to it failed.  It is guaranteed to NOT have the correct data.  The data MUST be re-constructed onto it by using parity and the other data disks.  For that reason, you should NOT delay in installing the parity disk.

2.5) I've installed monthly parity checks via cron. Are there any other maintenance-related procedures I should be aware of?

With no parity disk installed, that is not going to do anything.  You really need to install the parity disk and also perform a full parity check.  Only then will you know the array is working properly.

3) I installed unRAID-Web via unMENU but can't seem to access it via tower:89; I get 400 Bad Request, but I CAN access the web server at tower:2197; is there something I have to configure?

Absolutely no idea.

 

4) I recently learned that I should never use the "stop array" button in unMENU and always use the one in unRAID main menu instead. Are there any other similar "don't"'s that a newbie should know about?

Don't format a disk yourself.  Let unRAID do it for you.

5) I am running 4.7; I've been reading some legacy threads about increasing the read lookahead to increase read speeds, is this still necessary in current versions?

No.  But you can use the button on the user-scripts page to try alternate values to see if they help you.

6) I went through the list of packages in unMENU and a lot of them are very interesting and "nice to have", however would installing a ton of these packages onto cron jobs make the server noticeably slower?

Cron jobs only run at the time you specify.  Other than the parity check, most take seconds to run.

7) I read through the thread about 'go' scripts and was somewhat confused. http://lime-technology.com/forum/index.php?topic=5976.0 In a nutshell, what are some tweaks that are must have or highly recommended?

You should not be making changes to the go script based on other peoples needs.   Leave it as simple as possible.  Most of the tweeks are for special purposes.

 

Some people here love the Supermicro cards, whereas others opt for the ones from Monoprice. What's the difference, performance wise?

Probably nothing you'll ever notice.  I have neither, but both seem to have good reports.

9) I'm looking to automate, say, weekly or monthly backup of a small subset of important data to another location, either a computer, via web, or even Dropbox. What is possible with unRAID and what is the most practical solution?

Lots of solutions.   Look into crashplan, it seems to have a following.   Home-grown solutions using cron and rsync may also work depending on the target.

10) Is there anything similar to jDownloader that is supported by unRAID?

Don't know.  Don;t know what jDownloader does.

Lastly, I would also appreciate if there are anything you feel are "must-reads" for a newbie and wished someone told you when you first started using unRAID.

The whole "Configuration_Tutorial" wiki http://lime-technology.com/wiki/index.php?title=Configuration_Tutorial

For question three... I access unMenu at http://tower:8080 - did you try that port? It is the default...

 

Edit - never mind, wrong app.

8080 is the unMenu default. 80 is the unRAID default.

Very detailed, Joe, thanks!

 

The only reason why I haven't installed my parity drive yet is because I'm still in the process of migrating data and from what I've read, migrating with the parity drive will make the process much slower due to the parity calculations. I should be parity safe within the next two days though (including the full day it'll take to format the drive)

 

So from what I gather, the moment I add the parity disk and start the array up, it should automatically calculate parity and I will be 'safe' after that's completed. Afterwards, just do monthly parity checks and that should be enough to keep my array healthy?

 

BTW, I notice that the monthly parity check script doesn't send email updates as to whether or not something is wrong. I know you can use it along with the 'unRAID Status Alert sent hourly by e-mail' to get notified of issues, however I have received the hourly updates and they seem to be pretty complicated. If I just want to know whether or not the parity check succeeded or not, what should I look for in the status alert email? Is it possible to add some code that would just give a pass/fail with the parity check script via email?

 

For question three... I access unMenu at http://tower:8080 - did you try that port? It is the default...

 

Edit - never mind, wrong app.

 

=\

 

So nobody here uses unRAID-web? The GUI filesystem it had looked interesting and I wanted to try it, but can't access it

Hi Jon

 

Welcome to unRAID. I'm glad I and others on RFD could finally convince you to try it. Did you buy a licence? You'll need one to have three disks and a parity.

Some other advice:

 

Listen to whatever the experts like JoeL, BRiT, dgaschk among a few others (most are moderators) say. They've saved my ass and data a few times when I've been impatient!

So nobody here uses unRAID-web? The GUI filesystem it had looked interesting and I wanted to try it, but can't access it

It is just that it works for me.  No idea why it does not for you.  You'll need to look in its log file for clues (and I;ve no idea if lighttpd keeps one or has one enabled)

Immediately after building parity you need to run a parity check. Only if the parity check shows zero sync errors should you trust the array.

...and what are some good ways of monitoring drives so we know before they are about to fail? Are there any scripts that will do this and preferably send email notifications?

 

I was wondering about this recently myself... what is the best way to monitor and detect a drive about to fail?  (I have the parity check done weekly (with NOCORRECT option) so will notice quickly if the parity goes out of sync, but are there other ways to monitor and report to the user?

...and what are some good ways of monitoring drives so we know before they are about to fail? Are there any scripts that will do this and preferably send email notifications?

 

I was wondering about this recently myself... what is the best way to monitor and detect a drive about to fail?  (I have the parity check done weekly (with NOCORRECT option) so will notice quickly if the parity goes out of sync, but are there other ways to monitor and report to the user?

The best summary I've seen is the SMART report in unMENU.  It highlights failures that drives have encountered during writes, and will indicate drives that are starting to show problems.  But as far as I know, there is no way to "detect a drive about to fail".  That's like forecasting the weather.  You have your indications, but some drives can just die.

The unmenu package, "unRAID Status Alert sent hourly by e-mail" will inform you of an error condition hourly.

The unmenu package, "unRAID Status Alert sent hourly by e-mail" will inform you of an error condition hourly.

 

Do you know what the status alert emails look for when they determine if it needs the user's attention? Does it parse/process SMART reports?

The unmenu package, "unRAID Status Alert sent hourly by e-mail" will inform you of an error condition hourly.

 

Do you know what the status alert emails look for when they determine if it needs the user's attention?

yes. 

Does it parse/process SMART reports?

no.

The unmenu package, "unRAID Status Alert sent hourly by e-mail" will inform you of an error condition hourly.

 

 

 

Do you know what the status alert emails look for when they determine if it needs the user's attention? Does it parse/process SMART reports?

 

For me I've recieved things like

- Power failure on UPS Tower. Running on batteries.

- Power has returned on UPS Tower...

 

- Parity Check in process

Parity CHECK/RESYNC in progress,  96.9% complete,  est. finish in 16.8 minutes. Speed: 58142 kb/s.

Server Name: Tower

Server IP: 192.168.1.127

Date: Tue Mar  1 07:47:03 PST 2011

 

- Array not normal, which if I recall was because I lost a drive

 

- Array not running, because I shut it down tinkering and honestly walked away and forgot to restart it.

Just thought of something, how does spinning drives up and down affect the life of the drive? Will it cause a lot of wear and tear if the drives are being spun up/down often, because of a short spin down delay? I'm trying to look for the best balance between saving power and wearing of the drives.

 

For our usage as a media server, our usage is probably during 2-3 blocks during the day, which means with the delay set to a shorter time, our drives will be spun up/down up to 3 times a day.

The unmenu package, "unRAID Status Alert sent hourly by e-mail" will inform you of an error condition hourly.

 

Do you know what the status alert emails look for when they determine if it needs the user's attention?

yes. 

Does it parse/process SMART reports?

no.

I guess that just means, as the server administrator, we should probably be checking the SMART reports of the drives manually every so often?

Just thought of something, how does spinning drives up and down affect the life of the drive? Will it cause a lot of wear and tear if the drives are being spun up/down often, because of a short spin down delay? I'm trying to look for the best balance between saving power and wearing of the drives.

 

For our usage as a media server, our usage is probably during 2-3 blocks during the day, which means with the delay set to a shorter time, our drives will be spun up/down up to 3 times a day.

 

Historical performance is not an indication of future performance. New drive models are released more than yearly. There is no way of knowing this information. Starting and stopping a drive may not effect a drive any more than running it continuously. Manufacturers want enough of there drives to last for the warranty period that the cost of RMA's is offset by a cheaper manufacturing process. The drives are designed to make money for the vendor and satisfy the maximum number of end users. We have no idea how they go about this. This is a personal preferences (IMO mostly based on the cost of electricity).

Although facts are hard to come buy to answer questions like this, we can each have our own theories. I think excessive spin up / spin down would cause the drives to go through mote frequent heating up and cooling down cycles. Depending on your setup the temperature delta between spinning and not may be just a few degrees to 20C or more. IMO, sending drives through large temperature changes hurts their life expectancy.

You may be correct, although tolerance to thermal variation is definitely one of the design criteria.