Pablo L. Arturi | 14 Apr 2006 10:18

kernel error or hardware?

Hello people.

I have installed yesterday openvz, migrated a physical server to a vps and after a lot of help from the openvz
developers, I got it working excellent.

Today when I try to access the machine, it wasn't reachable in any way (remotely). What is seemed to be was
some fs problem, since when I tried to boot it, it always asked to run fsck, and it was telling all the time
there was problems with the partition /vz, among other things I will describe,

So well, I had to do like 20 server reboots, and in every reboot fsck started. After some time, it started
notmally, and everything working properly again.

I then started looking for logs, and found a lot of errors and strange (to me atleast) messages.

Googling some errors, in some cases there was some problem within the kernel, and in other cases there was a
hardware problem. That's what I coulnd't determine.

I am using 150GB sata disks, with a celeron processor 2.5 ghz, and 512 ram.

I will post my VPS config (the one I modified) and following all those strange log error messages. If anyone
has any idea about what could happened, it will be very appreciated.

Thanks in advance

[root <at> localhost log]# uname -a
Linux localhost.localdomain 2.6.8-022stab076.1 #1 Fri Apr 7 21:28:45 MSD 2006 i686 i686 i386 GNU/Linux

[root <at> localhost log]# cat /proc/user_beancounters
Version: 2.5
       uid  resource           held    maxheld    barrier      limit    failcnt
(Continue reading)

Aistis Zenkevicius | 14 Apr 2006 15:31
Picon

kernel error or hardware?

Hi

          I'm not the biggest expert, but it looks like your hdd lost
connection with controller, which led to broken ext3 journal.

Aistis

  _____  

From: users-bounces@...
[mailto:users-bounces@...] On
Behalf Of Pablo L. Arturi

Apr 13 01:32:52 localhost kernel: ata1: error occurred, port reset
Apr 13 01:32:52 localhost kernel: ata1: status=0x01 { Error }
Apr 13 01:32:52 localhost kernel: ata1: called with no error (01)!

Apr 13 04:13:14 localhost kernel: ata1: error occurred, port reset
Apr 13 04:13:14 localhost kernel: ata1: status=0x01 { Error }
Apr 13 04:13:14 localhost kernel: ata1: called with no error (01)!
Apr 13 04:15:01 localhost kernel: ata1: error occurred, port reset
Apr 13 04:15:01 localhost kernel: ata1: status=0x01 { Error }
Apr 13 04:15:01 localhost kernel: ata1: called with no error (01)!
Apr 13 04:15:05 localhost crond(pam_unix)[28711]: session opened for
user root by (uid=0)
Apr 13 04:15:05 localhost crond(pam_unix)[28712]: session opened for
user root by (uid=0)
Apr 13 04:15:06 localhost kernel: ata1: error occurred, port reset
Apr 13 04:15:06 localhost kernel: ata1: status=0x01 { Error }
Apr 13 04:15:06 localhost kernel: ata1: called with no error (01)!
(Continue reading)

Pablo L. Arturi | 15 Apr 2006 20:43

kernel error or hardware?

Thanks for your reply Aistis; do you think this is a hdd failure? It just happened again today, and I am going
to reproduce the same install on another server.

What do you think?

Thanks a lot
Pablo
  ----- Original Message ----- 
  From: Aistis Zenkevicius 
  To: users@... 
  Sent: Friday, April 14, 2006 2:31 PM
  Subject: RE: [Users] kernel error or hardware?

  Hi

            I'm not the biggest expert, but it looks like your hdd lost connection with controller, which led to broken
ext3 journal.

  Aistis

------------------------------------------------------------------------------

  From: users-bounces@...
[mailto:users-bounces@...] On Behalf Of Pablo L. Arturi

  Apr 13 01:32:52 localhost kernel: ata1: error occurred, port reset
  Apr 13 01:32:52 localhost kernel: ata1: status=0x01 { Error }
  Apr 13 01:32:52 localhost kernel: ata1: called with no error (01)!

  Apr 13 04:13:14 localhost kernel: ata1: error occurred, port reset
(Continue reading)

yok var | 16 Apr 2006 11:55
Picon
Favicon

[how-to] allocating disk quota / memory / cpu

Hi all,
 I am a bit confused after reading all stuff & forums about allocating ram / cpu and disk space for VPS's. Could
somebody please write down these commands with examples?

 Example of a scenario:
 VPSID: 100
 CPU: 101 MHZ
 RAM: 102 MB Ram
 HDD: 103 MB disk space

 Also, can we allocate resources for VPS's, over than host machine has? 

 Example of scenerio:

 HOST
 CPU: 1000 MHZ
 RAM: 1001 MB
 HDD: 1003 MB

 VPS 1
 CPU: 1000 MHZ
 RAM: 600 MB
 HDD: 800 MB

 VPS 2
 CPU: 1000 MHZ
 RAM: 600 MB
 HDD: 800 MB

 Is this possible?
(Continue reading)

Aleksandar | 21 Apr 2006 08:36

Problem with OpenVZ on Fedora Core 5

Hello,

i've just installed the openvz packages for fedora core 5. When I tried 
to run  vzpkgcache -f fedora-core-5 I got this:
# vzpkgcache -f fedora-core-5
Creating cache for fedora-core-5-i386-default OS template
ERROR: Can't start VPS 100002

I ran this with strace -f to see what might be the cause and I came 
across this:
[pid 10368] open("/usr/lib/locale/locale-archive", O_RDONLY|O_LARGEFILE) 
= 0
[pid 10368] fstat64(0, {st_mode=S_IFREG|0644, st_size=54062576, ...}) = 0
[pid 10368] mmap2(NULL, 2097152, PROT_READ, MAP_PRIVATE, 0, 0) = 0xb7d3d000
[pid 10368] close(0)                    = 0
[pid 10368] ioctl(0, SNDCTL_TMR_TIMEBASE or TCGETS, 0xbfc45af8) = -1 
EBADF (Bad file descriptor)
[pid 10368] unlink("/vz/tmp/XX0dwVO")   = -1 EISDIR (Is a directory)
[pid 10368] open(".", O_RDONLY|O_LARGEFILE) = 0
[pid 10368] fcntl64(0, F_DUPFD, 3)      = 11
[pid 10368] close(0)                    = 0

So something with ioctl() probably went wrong. Could this be because of 
the fedora core 5 kernel ? I've installed the fedora core 5 kernel from 
the openvz website and the vzctl and vzquota were complaining about 
missing vzkernel and vzkmod-quota but I've installed them with --nodeps 
(the vzquota module is there and was successfully loaded after "service 
vz start"). Also I've booted the openvz kernel:
# uname -r
2.6.16-1.2080_FC5.026test007
(Continue reading)

Aleksandar Kanchev | 21 Apr 2006 08:54
Picon
Favicon

Problem with OpenVZ on Fedora Core 5

Hello,

i've just installed the openvz packages for fedora core 5. When I tried 
to run  vzpkgcache -f fedora-core-5 I got this:
# vzpkgcache -f fedora-core-5
Creating cache for fedora-core-5-i386-default OS template
ERROR: Can't start VPS 100002

I ran this with strace -f to see what might be the cause and I came 
across this:
[pid 10368] open("/usr/lib/locale/locale-archive", O_RDONLY|O_LARGEFILE) = 0
[pid 10368] fstat64(0, {st_mode=S_IFREG|0644, st_size=54062576, ...}) = 0
[pid 10368] mmap2(NULL, 2097152, PROT_READ, MAP_PRIVATE, 0, 0) = 0xb7d3d000
[pid 10368] close(0)                    = 0
[pid 10368] ioctl(0, SNDCTL_TMR_TIMEBASE or TCGETS, 0xbfc45af8) = -1 
EBADF (Bad file descriptor)
[pid 10368] unlink("/vz/tmp/XX0dwVO")   = -1 EISDIR (Is a directory)
[pid 10368] open(".", O_RDONLY|O_LARGEFILE) = 0
[pid 10368] fcntl64(0, F_DUPFD, 3)      = 11
[pid 10368] close(0)                    = 0

So something with ioctl() probably went wrong. Could this be because of 
the fedora core 5 kernel ? I've installed the fedora core 5 kernel from 
the openvz website and the vzctl and vzquota were complaining about 
missing vzkernel and vzkmod-quota but I've installed them with --nodeps 
(the vzquota module is there and was successfully loaded after "service 
vz start"). Also I've booted the openvz kernel:
# uname -r
2.6.16-1.2080_FC5.026test007

(Continue reading)

Roberto Fichera | 21 Apr 2006 09:09
Picon
Favicon

Fedora Core 5 x86_64 precreated cache

Hi All,

Does anyone have a downloadable precreated FC5 x86_64
cache. I'm basically starting do configure a new x86_64 box
and since I'm unable to execute the vzpkgcache on such
box I would like to have a precreated one ;-)!

Thanks in advance.

Roberto Fichera. 

Kirill Korotaev | 24 Apr 2006 10:06
Picon

Problem with OpenVZ on Fedora Core 5

Hello,

which vzctl version do you have? try vzctl-3.0.0-2, as latest vzctl for 
i686 has one bug which leads to this err msg.
Will be updated today.

Thanks,
Kirill

> i've just installed the openvz packages for fedora core 5. When I tried 
> to run  vzpkgcache -f fedora-core-5 I got this:
> # vzpkgcache -f fedora-core-5
> Creating cache for fedora-core-5-i386-default OS template
> ERROR: Can't start VPS 100002
> 
> I ran this with strace -f to see what might be the cause and I came 
> across this:
> [pid 10368] open("/usr/lib/locale/locale-archive", O_RDONLY|O_LARGEFILE) 
> = 0
> [pid 10368] fstat64(0, {st_mode=S_IFREG|0644, st_size=54062576, ...}) = 0
> [pid 10368] mmap2(NULL, 2097152, PROT_READ, MAP_PRIVATE, 0, 0) = 0xb7d3d000
> [pid 10368] close(0)                    = 0
> [pid 10368] ioctl(0, SNDCTL_TMR_TIMEBASE or TCGETS, 0xbfc45af8) = -1 
> EBADF (Bad file descriptor)
> [pid 10368] unlink("/vz/tmp/XX0dwVO")   = -1 EISDIR (Is a directory)
> [pid 10368] open(".", O_RDONLY|O_LARGEFILE) = 0
> [pid 10368] fcntl64(0, F_DUPFD, 3)      = 11
> [pid 10368] close(0)                    = 0
> 
> So something with ioctl() probably went wrong. Could this be because of 
(Continue reading)

Kir Kolyshkin | 24 Apr 2006 12:07
Favicon

Problem with OpenVZ on Fedora Core 5

Alexander (and everybody else who's having the same problem),

We have just released vzctl-3.0.0-8 which should fix the issue with 
non-working vzpkgcache.

Changelog: http://openvz.org/news/updates/vzctl-3.0.0-8

Download: http://openvz.org/download/utils/vzctl/3.0.0-8

We apologize for any inconvenience.

Regards,
  Kir

Kirill Korotaev wrote:

> Hello,
>
> which vzctl version do you have? try vzctl-3.0.0-2, as latest vzctl 
> for i686 has one bug which leads to this err msg.
> Will be updated today.
>
> Thanks,
> Kirill
>
>> i've just installed the openvz packages for fedora core 5. When I 
>> tried to run  vzpkgcache -f fedora-core-5 I got this:
>> # vzpkgcache -f fedora-core-5
>> Creating cache for fedora-core-5-i386-default OS template
>> ERROR: Can't start VPS 100002
(Continue reading)

Isao Yamakawa | 25 Apr 2006 03:41

(no subject)


--

-- 
I Y <selfish@...>


Gmane