message "localhost cannot open /proc/meminfo"

Issues Related to Networking (Wired and Wireless)

Moderator: Moderators

Post Reply
User avatar
linuxfluesterer
Old Dear Hen
Posts: 870
Joined: Thu Sep 20, 2012 19:47
Location: Germany

message "localhost cannot open /proc/meminfo"

Post by linuxfluesterer » Sat Apr 28, 2018 16:34

Hallo guys.
Maybe this thread does not belong to network, but I assume to localhost phrase.
I have Sabayon 18.4, 64 bit with Kernel 4.15.17 and Plasma 5.12.4. While I was whatching tv with kaffeine (connected to a dvb-s2 receiver) I suddenly received the following message in an error window:

Code: Select all

message from localhost:
Cannot open '/proc/meminfo'!
The kernel needs to be compiled with support
for /proc file system enabled!
In my fstab, /proc, /sys and /run are dedicated to tmpfs (ram).

Goggle isn't my friend here. Any idea? Is it severe? I fear a complete crash could be the consequences.

Thank you in advance.

-Linuxfluesterer (I love KDE...)
Take away Facebook from me and let there be real people again...

User avatar
Fitzcarraldo
Sagely Hen
Posts: 8116
Joined: Sat Mar 10, 2007 5:40
Location: United Kingdom
Contact:

Re: message "localhost cannot open /proc/meminfo"

Post by Fitzcarraldo » Sat Apr 28, 2018 22:38

The Sabayon Linux 4.15.* kernel was compiled with support for /proc file system enabled (CONFIG_PROC_FS=y), as you can see: https://github.com/Sabayon/kernel/blob/ ... d64.config

So my pure guess is that Kaffeine is buffering/caching the video stream in RAM and reaching the limit/size of the buffer/cache.

Before you run Kaffeine, and while you run Kaffeine, check what /proc/meminfo contains:

Code: Select all

# cat /proc/meminfo
Also I suggest you check memory usage before, and while, Kaffeine is running:

Code: Select all

$ free --kibi --total
              total        used        free      shared  buff/cache   available
Mem:         997948      452092       94196        2692      451660      400700
Swap:       2097148         268     2096880
Total:      3095096      452360     2191076
And I suggest you check what is the memory use of tmpfs:

Code: Select all

$ sudo du -sh /run
928K    /run
$ df -h | grep run
none                              488M  916K  487M   1% /run
tmpfs                              98M   12K   98M   1% /run/user/1000
916K + 12K = 928K = amount of RAM in use by tmpfs. See: https://superuser.com/questions/542736/ ... amfs-tmpfs

Perhaps that will help you isolate the cause.

User avatar
linuxfluesterer
Old Dear Hen
Posts: 870
Joined: Thu Sep 20, 2012 19:47
Location: Germany

Re: message "localhost cannot open /proc/meminfo"

Post by linuxfluesterer » Sun Apr 29, 2018 11:12

Thank you, Fitz for this hint.
Exactly, I have 32GByte of Ram and I dedicated 9 GByte of Ram for tmpfs
So, when I enter both commands you suggested, I receive:

Code: Select all

# du -sh /run
1,3M    /run

# df -h | grep run
none             16G    1,3M   16G    1% /run
tmpfs           3,2G     12K  3,2G    1% /run/user/1000
I expected (that was my intention) to have enough RAM space to mount temporary directories into RAM (tmpfs) and also create a RAM disk of 9GByte also, created in fstab already. To be honest, I have no idea, which process or procedure dedicates 3.2 GBytes for tmpfs. In the following my /etc/fstab:

Code: Select all

UUID=a9cbd724-9978-41c0-87e3-384e817d50e2 /                       btrfs   defaults,ssd,noatime,space_cache,subvol=root   0 0
UUID=a9cbd724-9978-41c0-87e3-384e817d50e2 /Daten                  btrfs   defaults,ssd,noatime,space_cache,subvol=Daten  0 0
# UUID=75eeedbf-440a-4a0f-ad41-bebf95c7ccdc /boot                   ext4    defaults        1 2
UUID=EF84-F937                            /boot/efi               vfat    umask=0077,shortname=winnt 0 0
UUID=a9cbd724-9978-41c0-87e3-384e817d50e2 /home                   btrfs   defaults,ssd,noatime,space_cache,subvol=home   0 0
UUID=f58055bb-ff54-4826-aa6b-97568f6bf078 swap                    swap    defaults        0 0

/dev/sda4                                 /mnt/rootfs             btrfs   defaults,ssd,noatime,space_cache           0 0
tmpfs                                     /mnt/tmpfs              tmpfs   defaults,noatime,rw,mode=1777,size=9G,x-gvfs-show   0 0
devtmpfs                                  /dev                    devtmpfs defaults,noatime,rw          0       0 
tmpfs                                     /proc                   proc    defaults,noatime              0       0 
sysfs                                     /sys                    sysfs   nosuid,noexec,nodev           0       0
Don't wonder about excluded ext4 boot partition. In short I wanted to have all (linux) system files and data one ONE btrfs volume. This makes it much easier to recover a working snapshot after a broken upgrade.
So I copied all boot (etx4) contents to /boot (btrfs), changed my /etc/fstab, and executed the commands:

Code: Select all

grub2-mkconfig -o /boot/grub/grub2.cfg
grub2-install /dev/sda
Is there something, I can do to extend / dedicate some memory space for /run?

-Linuxfluesterer (I love KDE...)
Take away Facebook from me and let there be real people again...

Post Reply