[mehlvogel]

I again have a problem of a non-bootable Pi, this time I don't know caused from what, it was just sitting there. Let me fill out the template, I hope I get everything together.

Software
XBian version: Beta 2
XBMC version: I guess, latest
Overclock settings: Turbo

Hardware
Power supply rating: I am not sure what exactly was this means, but 1A is easily delivered to the Pi
RPi model (model A/B 256mb/512mb): 512mb / b
SD card size and make/type: 16 GB
Network (wireless or LAN): LAN
Connected devices (TV, USB, network storage, etc.): TV

Logile
Link to logfile(s): I don't know which ones would be relevant.

Problem description:
Freezes during boot.
Enabling "debug" and disabling "splash", the boot process gets stuck at "Starting udevtrigger".
After that, the blinking cursor at the bottom goes away and pressing Numlock does not trigger the light on the keyboard.
Booting into rescue mode, I had a look at the logs, but most of them are not even written at that time (I checked dmesg, messages, syslog, daemon.log and /run/up<something>)
The only "difference" or "anomality" I would guess relevant is that I have NFS shares mounted in my /etc/fstab.
I see the Pi requesting (and receiving) an IP from my DHCP, but after it freezes I am not able to ping (or connect using SSH).

How to reproduce:
At my place, by rebooting. With a fresh Pi, I have no idea

[particleman]

I am using a usb key for / but when I connect my powered usb drive I think Xbian becomes confused and does not know which drive holds the / so it will not boot unless my powered usb drive which holds my movies is unplugged any idea how to fix that?

[Koenkk]

I think your SD card is corrupted because of the turbo mode

[mk01]

(14th Dec, 2013 06:02 AM)particleman Wrote:  I am using a usb key for / but when I connect my powered usb drive I think Xbian becomes confused and does not know which drive holds the / so it will not boot unless my powered usb drive which holds my movies is unplugged any idea how to fix that?

yes, sure.

always we refer to root device as /dev/sda1 (example), but this is only to make things not too complicated.
but XBian accepts also UUID or LABEL parameters as root device.

that means you can instead of root=/dev/sda1 use root=UUID=421423q4-adsfsef-s14123 …

you can get UUID and LABEL by running "sudo blkid" in terminal with drives connected. if your copy wasn't created unluckily with dd UUID is unique and by using it you won't get conflict anymore.

(UUID or LABEL is resolved during boot into actual block device (sda,mmc …)).

---

(14th Dec, 2013 04:38 AM)mehlvogel Wrote:  The only "difference" or "anomality" I would guess relevant is that I have NFS shares mounted in my /etc/fstab.
I see the Pi requesting (and receiving) an IP from my DHCP, but after it freezes I am not able to ping (or connect using SSH).

post fstab and interfaces please and try with XBian clock (until I check the files).

[mehlvogel]

Hi, was the comment about the SD card directed at me? How can I check that? btrfs check says everything is fine (I have the SD card in my laptop now).

fstab: http://pastebin.com/ycPU1xBS
interfaces (I guess you meant /etc/network/interfaces): http://pastebin.com/sdvf0VnU

I am also not sure what you mean by XBian clock, but I guess you mean resetting my overclocking settings to Xbian defaults. I'll try that next.

Edit: (I few minutes down the road) It came back up after reverting the overclocking settings to the one in xbian beta2 images config.txt file. However, the NFS mounts did fail with the following message in dmesg:

Code:

[  269.016297] NFS: nfs4_discover_server_trunking unhandled error -512. Exiting with error EIO

Repeated quite a few times, the mount points had not been mounted after boot. Mounting them by hand hand (using sudo mount <mountpoint>) however worked, and after mounting all of them xbmc sprang to live again (I guess since these mountpoints are configured as library paths).

[mk01]

@mehlvogel

yes XBian clock = XBian OC setting (available through xbian-config for instance).

this looks again like the switch in preference nfs4 over nfs for mounting with mount -a . this happened with Debian 7.2 updates. nfs/nfs4 as filesystem type is no more relevant (nfs4 was depreciated) and vers=3/4 as mount option is recommended.

(this theory comes from my assumption your exports are nfs2/3 without fsid=0 root …)

if this is true please add "vers=3" to your mount options and try to reboot. if this is not the case please send post your /var/log/upstart/xbian-net.log .

[mehlvogel]

Thanks for your help, that seems to have fixed it. I rebootet a few times and it did mount everytime flawlessly.

Just out of curiosity: Do you happen to have an idea why changing the OC settings did fix the boot issue? Since the btrfs check said everything was fine, this strikes me as a bit odd. Or what am I missing?

[mk01]

@mehlvogel

this is because btrfs is checksumming data and metadata and checking with each read & write operation. so what happens is that if clock is too high and the bits get's messed up (doesn't matter if lets say during read or write) btrfs realizes that and remounts filesystem as readonly. you don't see this happening but as this happens system will fail to boot (because writes are essential during boot - for tmpfs creation, lock, status and socket files etc).

so for you us it looks like a fatal error (not booting) but in reality it saved you filesystem from corruption (because of RO no damaged data metadata gets written down). this is also the reason while since btrfs change we see so few reports on messed up filesystems. more it is like failing to start.

in contrast with ext4 for instance, corrupted structures (or 'only' data) gets written / read silently to the filesystem - clearly leading to disaster sooner or later - without indication - so you may think that fine, OC is good and stable. but this is not right.

from my experience OC like [ 900 ] [ 400 ] [ 450 ] [ 2 ] is quite stable and experience very little different from Turbo.

[Carpenter]

(13th Dec, 2013 10:01 PM)mk01 Wrote:  apt-get update
apt-get install --reinstall xbian-package-kernel

---

wrong package got into pool by mistake.

---

version will stay the same 1.3-6.3 but content will be right

I wish I could update because it says there is no space left on the device although df -h says something else:

Terminal

root@xbian:~# df -h
Filesystem Size Used Avail Use% Mounted on
rootfs 3.5G 2.0G 1.5G 59% /
/dev/mmcblk0p2 3.5G 2.0G 1.5G 59% /
devtmpfs 125M 4.0K 125M 1% /dev
none 25M 2.3M 23M 9% /run
/dev/mmcblk0p1 34M 20M 15M 57% /boot
/dev/mmcblk0p2 3.5G 2.0G 1.5G 59% /home
/dev/mmcblk0p2 3.5G 2.0G 1.5G 59% /lib/modules

[mk01]

(14th Dec, 2013 09:09 PM)Carpenter Wrote:  \

Terminal

root@xbian:~# df -h
Filesystem Size Used Avail Use% Mounted on
rootfs 3.5G 2.0G 1.5G 59% /
/dev/mmcblk0p2 3.5G 2.0G 1.5G 59% /
devtmpfs 125M 4.0K 125M 1% /dev
none 25M 2.3M 23M 9% /run
/dev/mmcblk0p1 34M 20M 15M 57% /boot
/dev/mmcblk0p2 3.5G 2.0G 1.5G 59% /home
/dev/mmcblk0p2 3.5G 2.0G 1.5G 59% /lib/modules

can you provide ssh for remote access?? I would be interested in that one case as it was reported few times but I could never "check" it myself.

what "btrfs fi df /" is saying ?

---

@Carpenter

copied text from https://btrfs.wiki.kernel.org/index.php/Problem_FAQ#I_get_.22No_space_left_on_device.22_errors.2C_but_df_says_I.27ve_got​_lots_of_space

Quote:I get "No space left on device" errors, but df says I've got lots of space

First, check how much space has been allocated on your filesystem:

Code:

$ sudo btrfs fi show
Label: 'media'  uuid: 3993e50e-a926-48a4-867f-36b53d924c35
    Total devices 1 FS bytes used 61.61GB
    devid    1 size 133.04GB used 133.04GB path /dev/sdf

Note that in this case, all of the devices (the only device) in the filesystem are fully utilised. This is your first clue.
Next, check how much of your metadata allocation has been used up:

Code:

$ sudo btrfs fi df /mount/point
Data: total=127.01GB, used=56.97GB
System, DUP: total=8.00MB, used=20.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=3.00GB, used=2.32GB
Metadata: total=8.00MB, used=0.00

Note that the Metadata used value is fairly close (75% or more) to the Metadata total value, but there's lots of Data space left. What has happened is that the filesystem has allocated all of the available space to either data or metadata, and then one of those has filled up (usually, it's the metadata space that does this). For now, a workaround is to run a partial balance:

Code:

$ sudo btrfs fi balance start -dusage=5 /mount/point

Note that there should be no space between the -d and the usage. This command will attempt to relocate data in empty or near-empty data chunks, allowing the space to be reclaimed and reassigned to metadata. More information is available elsewhere in this wiki, if you want to know what a balance does, or what options are available for the balance command.

[Carpenter]

(14th Dec, 2013 09:27 PM)mk01 Wrote:  

(14th Dec, 2013 09:09 PM)Carpenter Wrote:  \

Terminal

root@xbian:~# df -h
Filesystem Size Used Avail Use% Mounted on
rootfs 3.5G 2.0G 1.5G 59% /
/dev/mmcblk0p2 3.5G 2.0G 1.5G 59% /
devtmpfs 125M 4.0K 125M 1% /dev
none 25M 2.3M 23M 9% /run
/dev/mmcblk0p1 34M 20M 15M 57% /boot
/dev/mmcblk0p2 3.5G 2.0G 1.5G 59% /home
/dev/mmcblk0p2 3.5G 2.0G 1.5G 59% /lib/modules

can you provide ssh for remote access?? I would be interested in that one case as it was reported few times but I could never "check" it myself.

what "btrfs fi df /" is saying ?

---

@Carpenter

copied text from https://btrfs.wiki.kernel.org/index.php/Problem_FAQ#I_get_.22No_space_left_on_device.22_errors.2C_but_df_says_I.27ve_got​_lots_of_space

Quote:I get "No space left on device" errors, but df says I've got lots of space

First, check how much space has been allocated on your filesystem:

Code:

$ sudo btrfs fi show
Label: 'media'  uuid: 3993e50e-a926-48a4-867f-36b53d924c35
    Total devices 1 FS bytes used 61.61GB
    devid    1 size 133.04GB used 133.04GB path /dev/sdf

Note that in this case, all of the devices (the only device) in the filesystem are fully utilised. This is your first clue.
Next, check how much of your metadata allocation has been used up:

Code:

$ sudo btrfs fi df /mount/point
Data: total=127.01GB, used=56.97GB
System, DUP: total=8.00MB, used=20.00KB
System: total=4.00MB, used=0.00
Metadata, DUP: total=3.00GB, used=2.32GB
Metadata: total=8.00MB, used=0.00

Note that the Metadata used value is fairly close (75% or more) to the Metadata total value, but there's lots of Data space left. What has happened is that the filesystem has allocated all of the available space to either data or metadata, and then one of those has filled up (usually, it's the metadata space that does this). For now, a workaround is to run a partial balance:

Code:

$ sudo btrfs fi balance start -dusage=5 /mount/point

Note that there should be no space between the -d and the usage. This command will attempt to relocate data in empty or near-empty data chunks, allowing the space to be reclaimed and reassigned to metadata. More information is available elsewhere in this wiki, if you want to know what a balance does, or what options are available for the balance command.

@mk01: will send you the login details.

btrfs fi df /

Terminal

root@xbian:/tmp/btrfs-snap# btrfs fi df /
Data: total=3.08GB, used=1.66GB
System: total=4.00MB, used=16.00KB
Metadata: total=392.00MB, used=343.58MB

sudo btrfs fi show

Terminal

root@xbian:/tmp/btrfs-snap# sudo btrfs fi show
Label: 'xbian-beta2' uuid: 67901a92-640f-4087-96b2-7719a8367476
Total devices 1 FS bytes used 2.00GB
devid 1 size 3.46GB used 3.46GB path /dev/mmcblk0p2

Btrfs v0.20-rc1-358-g194aa4a

sudo btrfs fi balance start -dusage=5 /mount/point

Terminal

root@xbian:/tmp/btrfs-snap# sudo btrfs fi df /mount/point
ERROR: can't access to '/mount/point'

[mk01]

@Carpenter


sorry for "/mount/point" … as said I just re-copied the text.

in your case should be just "/". but according your "fi df /" it is the case with full metadata (or full on 75% and more) and last command should solve it.

---

your Beta2 was clean IMG from web or upgrade of beta1?

(according the Label it was IMG from web, clean, right ? )

[Carpenter]

(14th Dec, 2013 09:54 PM)mk01 Wrote:  @Carpenter


sorry for "/mount/point" … as said I just re-copied the text.

in your case should be just "/". but according your "fi df /" it is the case with full metadata (or full on 75% and more) and last command should solve it.

---

your Beta2 was clean IMG from web or upgrade of beta1?

(according the Label it was IMG from web, clean, right ? )

@mk01: it was a clean install from the web indeed

Terminal

sudo: unable to write to /var/lib/sudo/xbian/0: No space left on device
Data: total=3.08GB, used=1.66GB
System: total=4.00MB, used=16.00KB
Metadata: total=392.00MB, used=343.58MB

I don't think the last command works

Terminal

root@xbian:~# sudo btrfs fi balance start -dusage=5 /
Done, had to relocate 0 out of 15 chunks

[Skywatch]

Hi,

After an upgrade to Beta 2 I can no longer SSH in via any port except 22. I have always used another port and in past versions (1a3. 1a4. 1a5. 1b1.1) it has always worked. But now when I use the same putty preset it won't work.

I thought it was in issue with the key pair but it seems that on further investigation it's to do with the port being used.

If I log in on my usual port it won't have anything to do with it, but if I use port 22 it connects and authenticates OK.

I have checked /etc/smb/samba_config but that has the right port in it....and no sign of port 22.

I have also looked in the xbian ssh and smb config and can't find what is causing it to *only* look at port 22 now.

Any ideas?

cheers

skywatch

[mk01]

@Skywatch

ssh differently (Click to View)

(15th Dec, 2013 05:20 AM)Skywatch Wrote:  Hi,

After an upgrade to Beta 2 I can no longer SSH in via any port except 22. I have always used another port and in past versions (1a3. 1a4. 1a5. 1b1.1) it has always worked. But now when I use the same putty preset it won't work.

I thought it was in issue with the key pair but it seems that on further investigation it's to do with the port being used.

If I log in on my usual port it won't have anything to do with it, but if I use port 22 it connects and authenticates OK.

I have checked /etc/smb/samba_config but that has the right port in it....and no sign of port 22.

I have also looked in the xbian ssh and smb config and can't find what is causing it to *only* look at port 22 now.

Any ideas?

cheers

skywatch

search forum for post containing "ssh-nid"

---

his post, reply to freddie
http://forum.xbian.org/thread-1770-post-18223.html#pid18223