[vserver] Exploding Load in v2.3

From: Cryptronic <mail_at_cryptronic.de>
Date: Mon 18 Jan 2010 - 20:28:51 GMT
Message-ID: <4B54C483.60303@cryptronic.de>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi list,

we had problems on different kernel versions since vserver patch 2.3.

We had the problem, that load runs in a view seconds into great
dimensions:

top - 21:11:53 up 23 days, 5:55, 2 users, load average: 1092.35,
530.76, 212.88
Tasks: 5614 total, 3 running, 5531 sleeping, 0 stopped, 0 zombie
Cpu(s): 3.6%us, 8.2%sy, 0.0%ni, 83.2%id, 4.6%wa, 0.0%hi,
0.3%si, 0.0%st
Mem: 74376860k total, 73832132k used, 544728k free, 4835704k buffers
Swap: 140624968k total, 4019584k used, 136605384k free, 24147056k cached

As you can see "wa" is not very high.

At a load of 500 Apache and Co stop working and delivering webpages.

Load before it explodes is around 3 - 5.

On shell in host system you can work as at load levels like 0.50, but
vserver operations aren't possible: vserver XXX stop does not take any
affect.

dmesg is also empty.

We're using the following versions:

# vserver-info
Versions:
                   Kernel: 2.6.31.7-vs2.3.0.36.27
                   VS-API: 0x00020305
             util-vserver: 0.30.216-pre2864; Nov 29 2009, 00:39:14

Features:
                       CC: gcc, gcc (Debian 4.3.4-6) 4.3.4
                      CXX: g++, g++ (Debian 4.3.4-6) 4.3.4
                 CPPFLAGS: ''
                   CFLAGS: '-Wall -g -O2 -std=c99 -Wall -pedantic -W
- -funit-at-a-time'
                 CXXFLAGS: '-g -O2 -ansi -Wall -pedantic -W
- -fmessage-length=0 -funit-at-a-time'
               build/host: x86_64-pc-linux-gnu/x86_64-pc-linux-gnu
             Use dietlibc: yes
       Build C++ programs: yes
       Build C99 programs: yes
           Available APIs: v13,net,v21,v22,v23,netv2
            ext2fs Source: e2fsprogs
    syscall(2) invocation: alternative
      vserver(2) syscall#: 236/glibc
               crypto api: nss
          python bindings: no
   use library versioning: yes

Paths:
                   prefix: /usr
        sysconf-Directory: /etc
            cfg-Directory: /etc/vservers
         initrd-Directory: $(sysconfdir)/init.d
       pkgstate-Directory: /var/run/vservers
          vserver-Rootdir: /var/lib/vservers

At Kernel 2.6.29.6-vs2.3.0.36.14 this also happend. The hardware is
also not the thing, because this happens on different architectures.

To get the server back online, you have to reset the hardware, reboot
and co not doing anything.

Best regards
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.14 (GNU/Linux)

iEYEARECAAYFAktUxIIACgkQOBdlVlcPuhwnbACfauNJx2jiKTeCl/u/6KPw5m3G
yugAoMcI72qqRv1XDWa6wtarPrhc7WG7
=SxKC
-----END PGP SIGNATURE-----
Received on Mon Jan 18 20:29:21 2010

[Next/Previous Months] [Main vserver Project Homepage] [Howto Subscribe/Unsubscribe] [Paul Sladen's vserver stuff]
Generated on Mon 18 Jan 2010 - 20:29:27 GMT by hypermail 2.1.8