What to do when unRAID WebGUI can't be reached anymore?

[tillkrueger]

i can telnet into the my unRAID, but the webGUI is not coming up...i tried rebooting twice already, but no go...any pointers?

[Joe L.]

i can telnet into the my unRAID, but the webGUI is not coming up...i tried rebooting twice already, but no go...any pointers?

Same as always... attach a syslog to your next post.  Instructions in the wiki under "troubleshooting"

[tillkrueger]

if i generate a syslog to boot and execute the syslog in terminal, i only get this:

 

./syslog: line 1: Jul: command not found

./syslog: line 2: Jul: command not found

./syslog: line 3: syntax error near unexpected token `('

./syslog: line 3: `Jul  4 11:09:49 unRAID kernel: Linux version 2.6.37.6-unRAID (root@Develop) (gcc version 4.4.4 (GCC) ) #2 SMP Sat Jun 4 11:11:19 MDT 2011'

 

and generating the syslogtail, i get this:

 

root@unRAID:/boot# syslogtail.txt

./syslogtail.txt: line 1: syntax error near unexpected token `('

./syslogtail.txt: line 1: `Jul  4 11:09:49 unRAID kernel: ata14.00: 1953525168 sectors, multi 0: LBA48 NCQ (depth 31/32)'

 

the last thing i did last night before hitting the sack, was to execute the "New Permissions" script through the WebGUI (because i have run into endless permissions errors when trying to back up data from my other drives *to* the unRAID...in the morning the unRAID server was gone from the network, and i was unable to reach it through the browser (telnet is fine)...none of the shares show up in the finder, as if the server is switched off, but since telnet can reach it, it's obviously not.

 

did i do something bad by trying the "new permissions"? could that have caused it?

 

 

[tillkrueger]

so when the syslog says "line1", what script is it referring to? the go script?

the log is awfully short...did i not do this correctly?

 

since i have no access to any shared disks or user shares, it's difficult for me to get anything done regarding backup and data organization (which is what this entire week is all about).

[Joe L.]

if i generate a syslog to boot and execute the syslog in terminal, i only get this:

 

./syslog: line 1: Jul: command not found

./syslog: line 2: Jul: command not found

./syslog: line 3: syntax error near unexpected token `('

./syslog: line 3: `Jul  4 11:09:49 unRAID kernel: Linux version 2.6.37.6-unRAID (root@Develop) (gcc version 4.4.4 (GCC) ) #2 SMP Sat Jun 4 11:11:19 MDT 2011'

 

and generating the syslogtail, i get this:

 

root@unRAID:/boot# syslogtail.txt

./syslogtail.txt: line 1: syntax error near unexpected token `('

./syslogtail.txt: line 1: `Jul  4 11:09:49 unRAID kernel: ata14.00: 1953525168 sectors, multi 0: LBA48 NCQ (depth 31/32)'

 

the last thing i did last night before hitting the sack, was to execute the "New Permissions" script through the WebGUI (because i have run into endless permissions errors when trying to back up data from my other drives *to* the unRAID...in the morning the unRAID server was gone from the network, and i was unable to reach it through the browser (telnet is fine)...none of the shares show up in the finder, as if the server is switched off, but since telnet can reach it, it's obviously not.

 

did i do something bad by trying the "new permissions"? could that have caused it?

 

 

Both of those show you are not typing the command properly in the terminal. In both cases, it appears as if you are attempting to invoke the log file after capturing its contents. (and it does not contain commands to be executed)

 

What "exactly" are you typing in your attempt to capture the system log?

 

Joe L.

[tillkrueger]

oh, now i see what's probably happening here...you're right (as always) in that i just tried to invoke the syslog by typing "syslog" into the terminal (after typing "cp  /var/log/syslog  /boot" to generate it to the boot folder)...i first tried "print syslog" hoping it would spell it out in the terminal window itself so that i could copy it and paste it into the forum here...what is the proper syntax to print the syslog to the terminal (or get it into my computer)? since i don't have any connectivity other than through the terminal right now, that would be the most apparent way for me to get the proper info to this board, but i may be wrong...excuse my terminal noobiness/helplessness.

 

[Joe L.]

oh, now i see what's probably happening here...you're right (as always) in that i just tried to invoke the syslog by typing "syslog" into the terminal (after typing "cp  /var/log/syslog  /boot" to generate it to the boot folder)...i first tried "print syslog" hoping it would spell it out in the terminal window itself so that i could copy it and paste it into the forum here...what is the proper syntax to print the syslog to the terminal (or get it into my computer)? since i don't have any connectivity other than through the terminal right now, that would be the most apparent way for me to get the proper info to this board, but i may be wrong...excuse my terminal noobiness/helplessness.

 

As I said in my first post: Instructions in the wiki under "troubleshooting"

http://lime-technology.com/wiki/index.php?title=Troubleshooting#Capturing_your_syslog

[tillkrueger]

i know it must be frustrating sometimes, Joe, when people don't seem to listen (or read), but i thought i had read through all of the suggestions in the Wiki and did what i could with *only* being able to terminal into the Flash stick...since i wasn't able to see *any* of the disks, even in terminal, i wanted to avoid having to open the server (my USB stick is on the mobo)...but since i don't wanna be all-caps'ed by you, i did power down the server brutally (it won't shut down via the powerdown command) and removed the stick...

 

attached, please find the syslog...thanks for your patience!

syslog.txt

[Joe L.]

i know it must be frustrating sometimes, Joe, when people don't seem to listen (or read), but i thought i had read through all of the suggestions in the Wiki and did what i could with *only* being able to terminal into the Flash stick...since i wasn't able to see *any* of the disks, even in terminal, i wanted to avoid having to open the server (my USB stick is on the mobo)...but since i don't wanna be all-caps'ed by you, i did power down the server brutally (it won't shut down via the powerdown command) and removed the stick...

 

attached, please find the syslog...thanks for your patience!

 

There is nothing in your syslog that shows that the unRAID array was ever started...  or that the management interface was ever started.

 

I suspect you edited the config/go file to add in whatever you did to get subsonic in place, clobbered what was in the file prior to your change, and are left with no management screen since it is never being invoked.

 

You can take a look at the config/go file in any editor.

 

The top few lines should look exactly like this:

#!/bin/bash

# Start the Management Utility

/usr/local/sbin/emhttp &

 

If not sure what you need, post what you do have in the config/go file here in your next post.

 

To prove my point,  if you reboot your server you can log in and type:

nohup /usr/local/sbin/emhttp &

and odds are you'll then be able to see the management console.  (You'll still need to fix your config/go script for a permanent solution)

 

I understand you not wanting to take the server apart, but in this case, there was no other solution you could easily use.  For the short term, plug the USB drive into an outside port.

 

Joe L.

[tillkrueger]

and again, your intuition and knowledge rules supreme...and i feel silly for not being able to catch such basics...how the "/usr/local/sbin/emhttp &" got deleted from the top-section of the go script is beyond me...i certainly didn't touch it, but sure enough that was it...just starting up the server, aaaand bingo! well, one disk is showing as having problems, but that may just be a lose connection (happened before)...i'll deal with that tomorrow...bed-time here.

 

another night of restless dreams avoided...thank you, Joe.

i'll watch, read and learn.

[tillkrueger]

or so i thought...the array doesn't seem to wanna spin up now...i hit "Start" a couple of times (each time waiting about 5mins), even rebooted the server once, but still no go...so, here we go again...just in case something got mangled when installing all those packages within unmenu, here is my go script as it stands (not sure i did the cache_dirs includes/excludes right, but i doubt that's causing the current problem):

 

#!/bin/bash

# Start the Management Utility

/usr/local/sbin/emhttp &

 

/boot/boot/cache_dirs -w

-e "Downloads*"

-e "_BAK*"

-i "Media*"

-i "Movies*"

-i "Music*"

-i "Production*"

-i "Software*"

 

installpkg /boot/packages_global/powerdown-1.02-noarch-unRAID.tgz

installpkg /boot/packages_global/jre-6u20-i586-1.txz

installpkg /boot/packages/simpleFeatures-0.4-unraid-speeding_ant.tgz

 

/boot/scripts/subsonic_start.sh

/boot/scripts/shutdown_jobs_add.sh

 

sleep 30

blockdev --setra 2048 /dev/md1

done

 

cd /boot/packages && find . -name '*.auto_install' -type f -print | sort | xargs -n1 sh -c  unraid_notify start

 

disk 10 still shows as red (even after re-inserting the drive cage, which usually fixes that).

 

and i re-generated the syslog and attached it...

 

but now i really am gonna call it a night until i tackle it again tomorrow.

syslog.txt

[Joe L.]

All this needs to be removed, or put on the same line as the cache_dirs command:

-e "Downloads*"

-e "_BAK*"

-i "Media*"

-i "Movies*"

-i "Music*"

-i "Production*"

-i "Software*"

 

It should look like this:

cache_dirs -w -e "Downloads* -e "_BAK*" -i "Media*" -i "Movies*" -i "Music*" -i "Production*" -i "Software*"

 

These two lines need to be removed:

blockdev --setra 2048 /dev/md1

done

 

The sleep 30 must come BEFORE you invoke subsonic.  Move those lines after the sleep 30.

Joe L.

[tillkrueger]

had to get back up and try that...my go script now reads like this:

 

#!/bin/bash

# Start the Management Utility

/usr/local/sbin/emhttp &

 

/boot/boot/cache_dirs -w -e "Downloads*" -e "_BAK*" -i "Media*" -i "Movies*" -i "Music*" -i "Production*" -i "Software*"

 

installpkg /boot/packages_global/powerdown-1.02-noarch-unRAID.tgz

installpkg /boot/packages_global/jre-6u20-i586-1.txz

installpkg /boot/packages/simpleFeatures-0.4-unraid-speeding_ant.tgz

 

sleep 30

 

/boot/scripts/subsonic_start.sh

/boot/scripts/shutdown_jobs_add.sh

 

cd /boot/packages && find . -name '*.auto_install' -type f -print | sort | xargs -n1 sh -c

 

unraid_notify start

 

the server boots fine, but the array shows as not started with disk 10 having a red dot (have to inquire into that tomorrow)...anyway, i figured that i can start it with one bad disk in an un-protected state, and hit "Start"...after a bit the array says "Started" (in a red/orange state), but it doesn't say that parity is invalid, like usual, giving me the option to rebuild (maybe because it cannot see disk 10 as being available?)...when i click on "Shares", it gives me a warning message up top, saying "Array must be Started to view shares." and a couple of lines of mis-formatted text (probably due to some glitches in the CSS of the SimpleFeatures UI extension).

 

i'm attaching the newest syslog...

syslog.txt

[jinjorge]

running into the same problem where the webGUI is not reachable all over a sudden.  When the server is rebooted, before the parity check is completed, the WebGUI is accessible, but thereafter no dice

 

running 4.6rc5

 

any ideas?

 

Thanks

syslog-2011-07-04.txt

[tillkrueger]

hey...if i knew how to properly interpret a syslog, i would, but i'm kind of where you are, jinjorge, in learning as i go with this system...have you tried to telnet into the server, and are you able to connect with it? i guess first it would be good to determine whether it's simply a WebGUI issue, or whether it's a more base-level problem.

[tillkrueger]

so i now commented out all but the first and last instructions of the go script (i also know now that the syntax of my directory in/exclusions for the cache_dirs are wrong, but don't yet know how to properly name them):

 

#!/bin/bash

# Start the Management Utility

/usr/local/sbin/emhttp &

 

#!/ /boot/boot/cache_dirs -w -e "Downloads*" -e "_BAK*" -i "Media*" -i "Movies*" -i "Music*" -i "Production*" -i "Software*"

 

#!/ installpkg /boot/packages_global/powerdown-1.02-noarch-unRAID.tgz

#!/ installpkg /boot/packages_global/jre-6u20-i586-1.txz

#!/ installpkg /boot/packages/simpleFeatures-0.4-unraid-speeding_ant.tgz

 

#!/ sleep 30

 

#!/ /boot/scripts/subsonic_start.sh

#!/ /boot/scripts/shutdown_jobs_add.sh

 

cd /boot/packages && find . -name '*.auto_install' -type f -print | sort | xargs -n1 sh -c  unraid_notify start

 

but even with this minimal go script, i still get strange behavior of the default WebGUI...it looks like this (see attachment).

 

i don't get how this took such a turn for the worse when all was working so well until 2 days ago.

[tillkrueger]

what could be causing the WebGUI to be corrupted now, and be fully dysfunctional...all the user shares (and disk shares) are nowhere to be seen on the network, although the unRAID-SMB and unRAID (AFP) shares do show up (albeit empty)...what could be going on?

 

(no changes since my last syslog)

[Joe L.]

what could be causing the WebGUI to be corrupted now, and be fully dysfunctional...all the user shares (and disk shares) are nowhere to be seen on the network, although the unRAID-SMB and unRAID (AFP) shares do show up (albeit empty)...what could be going on?

 

(no changes since my last syslog)

I strongly suggest you get rid of the add-on lines in the "go" script until you are stable.

(walk before you run)

 

It appears now as if you are installing a package that is overwriting the "php" config and not enabling the short tags.

http://lime-technology.com/forum/index.php?topic=11540.msg123329#msg123329

[jinjorge]

hey...if i knew how to properly interpret a syslog, i would, but i'm kind of where you are, jinjorge, in learning as i go with this system...have you tried to telnet into the server, and are you able to connect with it? i guess first it would be good to determine whether it's simply a WebGUI issue, or whether it's a more base-level problem.

 

yes, I forgot to mention that I can access the server via telnet.

 

 

[tillkrueger]

Joe, the only remaining non-default line in there was the last one, no? you're probably correct in assuming that one of the auto-start packages might be the culprit...i took out the last line and it now looks like this:

 

#!/bin/bash

# Start the Management Utility

/usr/local/sbin/emhttp &

 

#!/ /boot/boot/cache_dirs -w

#!/ -e "Downloads*" -e "_BAK*" -i "Media*" -i "Movies*" -i "Music*" -i "Production*" -i "Software*"

 

#!/ installpkg /boot/packages_global/powerdown-1.02-noarch-unRAID.tgz

#!/ installpkg /boot/packages_global/jre-6u20-i586-1.txz

#!/ installpkg /boot/packages/simpleFeatures-0.4-unraid-speeding_ant.tgz

 

#!/ sleep 30

 

#!/ /boot/scripts/subsonic_start.sh

#!/ /boot/scripts/shutdown_jobs_add.sh

 

#!/ cd /boot/packages && find . -name '*.auto_install' -type f -print | sort | xargs -n1 sh -c unraid_notify start

 

btw, when i open up the go script in TextWrangler, the last statement ("unraid_notify start) always shows up on a separate line on its own, at the very end, but in TextWrangler i have it on one line...which is the correct way?

 

-------

 

and yup, the WebGUI is whole again...now i'll try to bring up the array and see whether it can get parity again...more soon.

[Joe L.]

Joe, the only remaining non-default line in there was the last one, no? you're probably correct in assuming that one of the auto-start packages might be the culprit...i took out the last line and it now looks like this:

 

#!/bin/bash

# Start the Management Utility

/usr/local/sbin/emhttp &

 

#!/ /boot/boot/cache_dirs -w

#!/ -e "Downloads*" -e "_BAK*" -i "Media*" -i "Movies*" -i "Music*" -i "Production*" -i "Software*"

 

#!/ installpkg /boot/packages_global/powerdown-1.02-noarch-unRAID.tgz

#!/ installpkg /boot/packages_global/jre-6u20-i586-1.txz

#!/ installpkg /boot/packages/simpleFeatures-0.4-unraid-speeding_ant.tgz

 

#!/ sleep 30

 

#!/ /boot/scripts/subsonic_start.sh

#!/ /boot/scripts/shutdown_jobs_add.sh

 

#!/ cd /boot/packages && find . -name '*.auto_install' -type f -print | sort | xargs -n1 sh -c unraid_notify start

 

btw, when i open up the go script in TextWrangler, the last statement ("unraid_notify start) always shows up on a separate line on its own, at the very end, but in TextWrangler i have it on one line...which is the correct way?

 

Do not use TextMangler. (or whatever it is named)

 

Learn some basic shell syntax.  For most commands, they go on a line of their own with all their arguments unless you use the line-continuation character as the very last character on a line to continue it on the next.

[lionelhutz]

TextWrangler (Textmangler lol) says it will save in a unix friendly enviroment so it ~should~ be usable. However, I have opened many acting up script files directly on the server (use Midnight Commander - type mc) and usually found I had to fix the file after I screwed it up with an external editor somehow.

 

 

jinjorge - you really need to start you own thread for help. You likely have a completely different problem and it's too confusing to try and solve 2 seperate issues in 1 thread.

 

Peter

[Ford Prefect]

btw, when i open up the go script in TextWrangler, the last statement ("unraid_notify start) always shows up on a separate line on its own, at the very end, but in TextWrangler i have it on one line...which is the correct way?

 

...could be some non-printable chars are still present in the file.

From a telnet session do this:

# cat -vet /boot/config/go

and check for "$", "^I" or "^M" characters in places, where they shouldn't belong.

 

The vanilla go script would look like this:

#!/bin/bash^M$
# Start the Management Utility^M$
/usr/local/sbin/emhttp &^M$

[tillkrueger]

k...i'll try to refrain from using TW then.

 

so the WebGUI is back up, and i was able to start the array...disk- and user-shares also shows up in my finder again...but the array cannot be rebuilt with parity right now (or it doesn't give me that option)...here is my unRAID main screen (see attached).

 

so, as i said, i don't see the usual option to rebuild the array and start parity check! i see that disk 10 is still red, but it's there and recognized...is it not going to (or *cannot*) rebuild until the red disk is green? just for good measure i'll attach the current syslog as well.

 

logically, since there is no valid parity right now, disk 10 being green would have to be necessary to build parity again, no?

[tillkrueger]

oops, the syslog didn't get attached...

syslog.txt