[Pkg-xen-devel] Bug#608988: xen-utils-4.0: please consider supporting remus

Wed Jan 5 07:50:42 UTC 2011

Package: xen-utils-4.0
Version: 4.0.1-1
Severity: wishlist

I did some experiments with the xen remus HA system. With

overlord3:~$ debdiff xen_4.0.1-1.dsc xen_4.0.1-1lindi1.dsc
diff -Nru xen-4.0.1/debian/rules.real xen-4.0.1/debian/rules.real

--- xen-4.0.1/debian/rules.real 2010-08-02 16:10:13.000000000 +0300
+++ xen-4.0.1/debian/rules.real 2011-01-05 00:32:41.000000000 +0200
@@ -159,6 +159,7 @@
        install -D -m644 debian/xen-utils.README.Debian $(PACKAGE_DIR)/usr/share/doc/$(PACKAGE_NAME)/README.Debian
        install -D -m644 debian/xen-utils-$(VERSION).lintian-overrides $(PACKAGE_DIR)/usr/share/lintian/overrides/$(PACKAGE_NAME)
        dh_install --sourcedir=$(DIR) usr/lib/xen-$(VERSION)
+       dh_install --sourcedir=$(DIR) usr/bin/remus usr/lib/xen-$(VERSION)/lib/python
        dh_pysupport -V$(shell pyversions -rv) /usr/lib/xen-$(VERSION)
        dh_strip
        dh_makeshlibs -V

overlord3:~$ cat /local/xen/lindi-exp1/config.pvops
name = 'lindi-exp1'
vcpus = '1'
memory = '256'
kernel = '/local/xen/lindi-exp1/vmlinuz-2.6.32-5-amd64'
ramdisk = '/local/xen/lindi-exp1/initrd.img-2.6.32-5-amd64'
nfs_server = '10.7.2.17'
nfs_root = '/tftpboot/lindi-exp1'
extra = 'root=/dev/nfs ip=dhcp'
vif = [ 'mac=52:54:00:12:35:09' ]
cpuid = [ '0:eax=0x3,ebx=0x0,ecx=0x0,edx=0x0',
'1:eax=0x0f61,ecx=xxxxxxxx0xx00xxxxxxxxx0xxxxxxxxx,edx=xxx0xxxxxxxxxxxxxxxxxxxxxxxxxxxx',
'0x80000000:eax=0x80000004,ebx=0x0,ecx=0x0,edx=0x0',
'0x80000001:eax=0x0f61,ecx=xxxxxxxxxxxxxxxxxx0000000000000x,edx=00xx000xx0xxx0xxxxxxxxxxxxxxxxxx' ]
# up to leaf 0000_0003/8000_0004
# family 15 model 6 stepping 1 (Intel P4 Prescott, AMD K8)
# disable POPCNT, SSE4.[12], SSSE3
# disable HTT
# disable CMPLEGACY, SVM, EXTAPIC, ALTMOVCR8, ABM, SSE4a, MisAlignSSE,
#         3DNOWPrefetch, OSVW, IBS, SSE5, SKINIT, WDT
# disable 3DNOW, 3DNOWEXT, RDTSCP, Page1GB, FFXSR, MMXExt, MP
#
# Source: Cross-Vendor Migration
# AMD Operating Systems Research Center
# Uwe Dannowski       Andre Przywara
#   {Firstname.Lastname}@amd.com
# 2010 Advanced Micro Devices, Inc. All rights reserved.

overlord3:~$ dmesg | head -n3
[    0.000000] Initializing cgroup subsys cpuset
[    0.000000] Initializing cgroup subsys cpu
[    0.000000] Linux version 2.6.32-5-xen-amd64 (Debian 2.6.32-29) (ben at decadent.org.uk) (gcc version 4.3.5 (Debian 4.3.5-4) ) #1 SMP Fri Dec 10 17:41:50 UTC 2010

overlord3:~$ grep -v "^#" /etc/xen/xend-config.sxp|grep -v "^$"
(network-script network-bridge)
(vif-script vif-bridge)
(dom0-min-mem 512)
(enable-dom0-ballooning yes)
(total_available_memory 0)
(dom0-cpus 0)
(vncpasswd '')
(xend-relocation-server yes)

running

/usr/lib/xen-4.0/lib/python/remus --no-net lindi-exp1 10.7.2.11

creates a paused version of lindi-exp1 successfully on 10.7.2.11
(overlord1) and starts updating it with output like

...
PROF: resumed at 1294213416.981778
PROF: flushed memory at 1294213416.988018
PROF: suspending at 1294213417.180439
SUSPEND shinfo 000699f7
delta 1ms, dom0 100%, target 0%, sent 0Mb/s, dirtied 6225Mb/s 190 pages
 3220: sent 179, skipped 0, delta 1ms, dom0 0%, target 0%, sent 5865Mb/s, dirtied 5865Mb/s 179 pages
Total pages sent= 666245 (10.17x)
(of which 0 were fixups)
All memory is saved
PROF: resumed at 1294213417.183659
PROF: flushed memory at 1294213417.189961
PROF: suspending at 1294213417.382440
SUSPEND shinfo 000699f7
delta 1ms, dom0 100%, target 0%, sent 0Mb/s, dirtied 5865Mb/s 179 pages
 3221: sent 183, skipped 0, delta 1ms, dom0 0%, target 0%, sent 5996Mb/s, dirtied 5996Mb/s 183 pages
Total pages sent= 666428 (10.17x)
(of which 0 were fixups)
All memory is saved
PROF: resumed at 1294213417.385560
...

If I 

overlord3:~$ sudo xm destroy lindi-exp1

then the domU is unpaused on overlord1 and works just great:

overlord3:~$ ping lindi-exp1
PING lindi-exp1.lan (10.7.6.20) 56(84) bytes of data.
64 bytes from lindi-exp1.lan (10.7.6.20): icmp_req=1 ttl=64 time=2.60 ms
^C
--- lindi-exp1.lan ping statistics ---
1 packets transmitted, 1 received, 0% packet loss, time 0ms
rtt min/avg/max/mdev = 2.601/2.601/2.601/0.000 ms


Have you considered shipping the remus executable so that people could
at least experiment with remus even if it is not considered stable yet?

(Currently my largest problem is that lindi-exp1's networking does not
work when remus is started, it only works before remus and after xm
destroy. Is this because of --no-net? I thought it would only skip
extra buffering but still allow it to work...)


-- System Information:
Debian Release: squeeze/sid
  APT prefers testing
  APT policy: (500, 'testing')
Architecture: amd64 (x86_64)

Kernel: Linux 2.6.32-5-xen-amd64 (SMP w/6 CPU cores)
Locale: LANG=C, LC_CTYPE=fi_FI (charmap=ISO-8859-1)
Shell: /bin/sh linked to /bin/dash

Versions of packages xen-utils-4.0 depends on:
ii  e2fslibs                1.41.12-2        ext2/ext3/ext4 file system librari
ii  iproute                 20100519-3       networking and traffic control too
ii  libbz2-1.0              1.0.5-6          high-quality block-sorting file co
ii  libc6                   2.11.2-7         Embedded GNU C Library: Shared lib
ii  libncurses5             5.7+20100313-4   shared libraries for terminal hand
ii  libxenstore3.0          4.0.1-1lindi1    Xenstore communications library fo
ii  python-support          1.0.10           automated rebuilding support for P
ii  python2.5               2.5.5-11         An interactive high-level object-o
ii  udev                    164-3            /dev/ and hotplug management daemo
ii  xen-utils-common        4.0.0-1          XEN administrative tools - common 
ii  zlib1g                  1:1.2.3.4.dfsg-3 compression library - runtime

Versions of packages xen-utils-4.0 recommends:
ii  bridge-utils               1.4-5         Utilities for configuring the Linu
ii  xen-hypervisor-4.0-amd64 [ 4.0.1-1lindi1 The Xen Hypervisor on AMD64

Versions of packages xen-utils-4.0 suggests:
pn  xen-docs-4.0                  <none>     (no description available)

-- no debconf information



