Swap no longer working

SAFAD · December 9, 2011, 9:16am

Hi There,
Since couple of days My Opensuse 12.1 Box started to lag when I heavily load it, it used to work fine but now am not sure why the memory gets 100% loaded and it stops responding.
I’ve investigated a bit the problem, and it seems new disks are being added to load the RAM as file system (tmpfs), never had them and I don’t know how to disable them now.
Apparently these file systems are used when swap is not working to make sure Linux won’t crash when its out of memory, but what got it disabled and how to get it back ?
My swap space can be found in /dev/sda6, when I tried to re-enable it :

linux-rgws:/home/safad # swapon /dev/sda6
swapon: /dev/sda6: swapon failed: Device or resource busy

Tried to check if its already mounted as swap :

linux-rgws:/home/safad # swapon -s
Filename				Type		Size	         Used	   Priority
/dev/sda6                          partition	10486780	  252264	    0

something is not correct and I want my Box back as it was :’(
Best Regards

caf4926 · December 9, 2011, 9:20am

you should post for us the result of

cat /etc/fstab

su -
fdisk -l

SAFAD · December 9, 2011, 9:28am

yes Sorry about that :



linux-rgws:/home/safad # cat /etc/fstab
/dev/disk/by-id/ata-Hitachi_HDS5C1050CLA382_JC0550JQ0BL5GH-part5 /home                ext4       defaults              1 2
/dev/disk/by-id/ata-Hitachi_HDS5C1050CLA382_JC0550JQ0BL5GH-part1 /                    ext4       noatime,nodiratime,barrier=1,acl,user_xattr        1 1
proc                 /proc                proc       defaults              0 0
sysfs                /sys                 sysfs      noauto                0 0
debugfs              /sys/kernel/debug    debugfs    noauto                0 0
usbfs                /proc/bus/usb        usbfs      noauto                0 0
devpts               /dev/pts             devpts     mode=0620,gid=5       0 0
/dev/disk/by-id/ata-Hitachi_HDS5C1050CLA382_JC0550JQ0BL5GH-part6 swap                 swap       defaults              0 0
tmpfs                /tmp                 tmpfs      defaults              0 0


linux-rgws:/home/safad # fdisk -l

Disk /dev/sda: 500.1 GB, 500107862016 bytes
255 heads, 63 sectors/track, 60801 cylinders, total 976773168 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0xe919e919

   Device Boot      Start         End      Blocks   Id  System
/dev/sda1   *        2048   156459007    78228480   83  Linux
/dev/sda2       156461054   976752639   410145793    5  Extended
/dev/sda5       160366592   579791611   209712510   83  Linux
/dev/sda6       579792896   600766463    10486784   82  Linux swap / Solaris

caf4926 · December 9, 2011, 9:37am

/dev/disk/by-id/ata-Hitachi_HDS5C1050CLA382_JC0550JQ0BL5GH-part6 swap                 swap       defaults              0 0
/dev/disk/by-id/ata-Hitachi_HDS5C1050CLA382_JC0550JQ0BL5GH-part5 /home                ext4       defaults              1 2
/dev/disk/by-id/ata-Hitachi_HDS5C1050CLA382_JC0550JQ0BL5GH-part1 /                    ext4       noatime,nodiratime,barrier=1,acl,user_xattr        1 1
proc                 /proc                proc       defaults              0 0
sysfs                /sys                 sysfs      noauto                0 0
debugfs              /sys/kernel/debug    debugfs    noauto                0 0
usbfs                /proc/bus/usb        usbfs      noauto                0 0
devpts               /dev/pts             devpts     mode=0620,gid=5       0 0
tmpfs                /tmp                 tmpfs      defaults              0 0

Have you been messing with fstab? Because normally swap is the first entry as shown.
I doubt that makes a difference though, not sure.
Is there anything you have been fiddling with?

Run ‘top’ in a terminal to determine system loads

ken_yap · December 9, 2011, 9:54am

SAFAD:

My swap space can be found in /dev/sda6, when I tried to re-enable it :
linux-rgws:/home/safad # swapon /dev/sda6
swapon: /dev/sda6: swapon failed: Device or resource busy
Tried to check if its already mounted as swap :
linux-rgws:/home/safad # swapon -s
Filename				Type		Size	         Used	   Priority
/dev/sda6                          partition	10486780	  252264	    0

This is what you would expect to see if swap is active. What makes you think it’s not?

SAFAD · December 9, 2011, 10:09am

Well, I’ve tried to Increase the performance but am sure nothing has being changes in fstab, nor anything related to Swap, just disabled some services I don’t use and modifications to software I use all the time,
Should i Re-write fstab like you did above ?

caf4926 · December 9, 2011, 10:11am

First show us top running in a terminal

caf4926 · December 9, 2011, 10:14am

see this example
SUSE Paste

SAFAD · December 9, 2011, 10:47am

top - 10:07:44 up 2:25, 3 users, load average: 0.24, 0.28, 0.31 Tasks: 179 t - Pastebin.com

this is when am only doing some small coding, PHP edits and such things
but when I start openGL simmulations, Play HoN and start Memory Intensive application
Memory Loads 100% same to CPU,
it never did this, Just the last week when TMPFS appeared TçT

caf4926 · December 9, 2011, 11:16am

You have zombie processes
Probably chrome
I saw this recently

ken_yap · December 9, 2011, 11:30am

I would trust that unfounded guess as much as your demonstrably mistaken assertion that swap wasn’t active. Better collect more evidence before jumping to conclusions.

caf4926 · December 9, 2011, 11:39am

I just love the English language.
I had to read that twice and I’m an Englishman with a moderate grasp of the language…
If only we could run it in a terminal and send ‘swap’ ballistic

robin_listas · December 9, 2011, 12:13pm

On 2011-12-09 09:16, SAFAD wrote:

> I’ve investigated a bit the problem, and it seems new disks are being
> added to load the RAM as file system (tmpfs), never had them and I don’t
> know how to disable them now.

Not possible to disable. They are in fact only one disk, and the files
there are temporary and small. You can check with “df”.

See “http://lwn.net/Articles/436012/” for the long explanation.

–
Cheers / Saludos,

Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)

ken_yap · December 9, 2011, 12:22pm

My goal is one day to be able to launch into a blistering denunciation of a deserving retard and hear him thank me for the compliments afterwards. >:) I’m currently studying the sayings of Captain Haddock.

lwfinger · December 9, 2011, 5:17pm

On 12/09/2011 05:13 AM, Carlos E. R. wrote:
> On 2011-12-09 09:16, SAFAD wrote:
>
>> I’ve investigated a bit the problem, and it seems new disks are being
>> added to load the RAM as file system (tmpfs), never had them and I don’t
>> know how to disable them now.
>
> Not possible to disable. They are in fact only one disk, and the files
> there are temporary and small. You can check with “df”.
>
> See “http://lwn.net/Articles/436012/” for the long explanation.

Note: tmpfs has always been there. The change is the way it shows in the df list.

Please post the output of ‘free’ when you are having the “problem”.

DenverD · December 12, 2011, 2:25pm

On 12/09/2011 10:16 AM, SAFAD wrote:
> I’ve tried to Increase the performance

the GOOD news: your attempts to increase performance didn’t cause the
use of tmpfs (they have been there since the day you installed 12.1)

more GOOD news: your swap is working

even more GOOD news: it is common for very high levels of memory to be
in use (since RAM is much faster than the fastest hard disks, unused
memory is wasted memory)

probable bad news: in my opinion the cause of your “My Opensuse 12.1 Box
started to lag” is likely to be something you did when you “tried to
Increase the performance”…suggest you stop investigating swap, tmpfs
and memory usage (which all appear ok to me) and UNDO your performance
increasing actions until your box stops the lag…

–
DD http://tinyurl.com/DD-Caveat
openSUSE®, the “German Engineered Automobiles” of operating systems!

robin_listas · December 12, 2011, 2:43pm

On 2011-12-09 17:17, Larry Finger wrote:
> On 12/09/2011 05:13 AM, Carlos E. R. wrote:

> Note: tmpfs has always been there. The change is the way it shows in the df
> list.

Not really. The /run directory is new. Previously they were using space in /dev

–
Cheers / Saludos,

Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)

DenverD · December 13, 2011, 10:15am

On 12/12/2011 02:43 PM, Carlos E. R. wrote:
> On 2011-12-09 17:17, Larry Finger wrote:
>> On 12/09/2011 05:13 AM, Carlos E. R. wrote:
>
>
>> Note: tmpfs has always been there. The change is the way it shows in the df
>> list.
>
> Not really. The /run directory is new. Previously they were using space in /dev

these “problems” exist in an install of 12.1, and since the day 12.1 was
installed on the “problem” syatem, at least, tmpfs have always been there!!

no doubt about that.

–
DD
openSUSE®, the “German Engineered Automobiles” of operating systems!

robin_listas · December 13, 2011, 4:43pm

On 2011-12-13 10:15, DenverD wrote:
> On 12/12/2011 02:43 PM, Carlos E. R. wrote:

> these “problems” exist in an install of 12.1, and since the day 12.1 was
> installed on the “problem” syatem, at least, tmpfs have always been there!!
>
> no doubt about that.

Yes, the tmpfs has been in use for a long time, but 12.1 has a new
directory at /run, bind mounted to /var/lock, /var/run and /media. This is
new, by design.

You really should read the article at <http://lwn.net/Articles/436012/>, it
is very interesting.

–
Cheers / Saludos,

Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)

DenverD · December 13, 2011, 6:05pm

> very interesting.
thanks for the pointer…

–
DD