Bug#629253: linux-image-2.6.32-5-amd64: Squeeze amd64 PV DOMU live migration fails
Hi Bastian, thank you for your quick answer. I split this email in 2 parts. Things belonging to the bug I send to 629...@bugs.debian.org. The answers to the rest I sent to you. . . . Sometimes lm works but after migration machine is not reachable through network anymore. More information. xm network-list --long $domain before and after. Informations about the network setup. Kernel log from the dom0. I will do this this tests, but I have a lot of projects with higher priority at the moment. So please give me some time for that. Mostly lm crashes during migration. Messages? see above. . . . Kind regards Dipl.-Ing.(FH) Holger Fischer -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4deb715f.9020...@web.de
Bug#629253: linux-image-2.6.32-5-amd64: Squeeze amd64 PV DOMU live migration fails
Hi, can't provide test results, but other people having same problem and a bit more info. http://forums.citrix.com/message.jspa?messageID=1533046 Kind regards Dipl.-Ing.(FH) Holger Fischer -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4deb7a45.9070...@web.de
Bug#629253: linux-image-2.6.32-5-amd64: Squeeze amd64 PV DOMU live migration fails
Hi Basti, dear maintainers, Google debian squeeze live migration xenserver or debian squeeze \live migration\ citrix. The first results are not bad. Read them through and you get an idea (a few people have ideas how to fix). I tried googling for squeeze domu live migration brought me too much results for Squeeze as DOM0. Possibly debian squeeze live migration \xen cloud platform\ or ...\xcp\ or ...\oraclevm\ or ...\sles\ gives you some good results too. f.e. http://forums.citrix.com/thread.jspa?threadID=281439tstart=0 - post 13 Using 2.6.38 from backports is not an option. I'm not allowed to use an unmaintained kernel. Can remember I tested 2.6.38 and remind syslog-ng was not working/segfaulting. I believe if you test live migration yourself on your test HW with Squeeze DOMU (amd64) on a DOM0 of your choice, you'll get similar results. Kind regards Dipl.-Ing.(FH) Holger Fischer -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4deb82ff.6070...@web.de
Bug#629253: linux-image-2.6.32-5-amd64: Squeeze amd64 PV DOMU live migration fails
Package: linux-2.6 Version: 2.6.32-34 Severity: normal Hi, At first I must point out: For me Squeeze is far away from being the best Debian ever. I'm really disappointed from the quality of Debian Squeeze as server OS. I began testing when Squeeze became stable. Too many problems. Much too much time spent on testing, finding errors and fixing/workaround (my boss really loves me now for the work that did not succeed during the last few months). I don't know where to point this problems out at a higher position, because they are more general? Is there a contact to the Debian management where I can ask for more stability and completeness of Debian than for new features (possibly fine for a desktop only distri)? Answering this question would be nice! Our business must have all security fixes within 3 months after release. Also we have 24/7 web services, servers mostly clustered, where every minute downtime really costs money (if downtime is too long it costs existence). We can't stay on Lenny for very long. EOL is coming near. After about half a year Squeeze stable, it is far away from being ready for our production servers. That's why we decided for now to go with a RHEL clone for our new installed productive servers. I will keep an eye on Debian. Once the quality comes backwho knows. A few lines later I describe the problems I had testing Squeeze as XEN DOM0. My install system for Windows (OPSI) our Solaris servers (Jumpstart) and RHEL based distros (Kickstart) still resides on Debian Squeeze. Moving them to another platform is a lot of work. That's why it would be still nice to have a live migrateable Debian Squeeze. Here we go: Sometimes lm works but after migration machine is not reachable through network anymore. Somebody mentioned a cronjob with traffic on vnet-device would help keeping machine reachable through network. Mostly lm crashes during migration. I tested all versions from 2.6.32-30 -- -34. Dom0 is: Choose on - I tested: XCP 1.0 OpenSuse 11.3 with Xen 4.0.1 Debian Squeeze (4.0.1) Squeeze worked most bad as DOM0 and should not be advertised as a working enterprise solution for XEN virtualization: - crashes our dell r7xx servers when using multipath + iscsi or only iscsi - live migration (with a working os's(2008R2 f.e.) as DOMU ) works only 5-6 times then it crashes. f.e. scripted live migration of a 2008R2 and a Lenny DOMU (at the same time) between 2 nodes worked 2 days (DOMUs reachable all the time) with XCP 1.0 (about 230 live migrations each DOMU), then we stopped testing - xend freezes - random reboots of DOM0 due to other reasons that could not be located - lots of scaring error messages when uptime grows but back to live migration problem of Debian Squeeze DOMU's: DomU's are installed with FAI. Lenny with a similar install works just fine when: /proc/cmdline: ... clocksource=jiffies and /etc/sysctl.conf: ...\nxen_independent_wallclock=1\n Live Migration works also fine with Win2008R2 Winxpsp3 ncp 3.0.1 RHEL(PUIAS clone) 5.6 RHEL(PUIAS clone) 6.1 I assume the problem is the Squeeze kernel. -- Package-specific info: ** Version: Linux version 2.6.32-5-amd64 (Debian 2.6.32-34) ** Command line: root=/dev/mapper/vg01-root ro console=tty0 console=hvc0 ** Not tainted ** Model information not available ** Loaded modules: Module Size Used by nls_utf81208 0 isofs 27480 0 loop 11799 0 autofs420805 8 nfsd 254782 13 exportfs3170 1 nfsd nfs 241114 1 lockd 57603 2 nfsd,nfs fscache29834 1 nfs nfs_acl 2031 2 nfsd,nfs auth_rpcgss33476 2 nfsd,nfs sunrpc161541 15 nfsd,nfs,lockd,nfs_acl,auth_rpcgss btrfs 375670 1 zlib_deflate 17746 1 btrfs crc32c 2560 1 libcrc32c 1074 1 btrfs ext3 106518 1 jbd37085 1 ext3 psmouse49937 0 usbhid 33292 0 hid63257 1 usbhid uhci_hcd 18521 0 ohci_hcd 19343 0 ehci_hcd 32081 0 usbcore 122386 4 usbhid,uhci_hcd,ohci_hcd,ehci_hcd nls_base6377 3 nls_utf8,isofs,usbcore dm_crypt 10664 0 snd_pcm60503 0 snd_timer 15582 1 snd_pcm snd46446 2 snd_pcm,snd_timer soundcore 4598 1 snd snd_page_alloc 6249 1 snd_pcm pcspkr 1699 0 joydev 8459 0 evdev 7352 0 ext4 289033 1 mbcache 5050 2 ext3,ext4 jbd2 67015 1 ext4 crc16 1319 1 ext4 dm_mod 53706 10 dm_crypt raid10 17809 0 raid45644500 0
Bug#593276: initramfstools =0.94 not compatible with lenny?
package: initramfs-tools
version: 0.97
Hi there,
last weekend I was testing something with fai, initramfs-tools and
live-initramfs, because I had trouble with netbooting a via vb8001 mb
with a second pcie e1000e card.
My current solution for initramfs-tools is:
- take source code of initramfs-tools_0.93.4-grml02 from
http://deb.grml.org/pool/main/i/initramfs-tools
- revert commit 836e5cf6ddb30757365c2b879b96958c250dcd7b - mkinitramfs:
only copy modprobe conf files
(which was reverted in commit 60e66ab6f7e799ac99e051fd8877f5b54758f454
that's not included in initramfs-tools_0.93.4-grml02),
because it breaks compatibility with lenny.
- compile and make a new version for my local repo
Here's a diff of the source codes of initramfs-tools_0.93.4-grml02 and
my local version:
diff -Naurp initramfs-tools-0.93.4-grml02/debian/changelog
initramfs-tools-0.93.4-grml02~hbpo02/debian/changelog
--- initramfs-tools-0.93.4-grml02/debian/changelog2010-08-16
23:03:01.0 +0200
+++ initramfs-tools-0.93.4-grml02~hbpo02/debian/changelog2010-08-16
23:03:33.0 +0200
@@ -1,3 +1,16 @@
+initramfs-tools (0.93.4-grml02~hbpo02) lenny; urgency=low
+
+ * revert copying only /etc/modprobe.d/*.conf to *
+as it breaks compatibility with lenny
+
+ -- hoonet local system user hoou...@hpedebsv19.hpe.hoonet.org Sun,
15 Aug 2010 19:06:42 +0200
+
+initramfs-tools (0.93.4-grml02~hbpo01) lenny; urgency=low
+
+ * rebuild for lenny
+
+ -- hoonet local system user hoou...@hpedebsv19.hpe.hoonet.org Thu,
20 May 2010 15:57:00 +0200
+
initramfs-tools (0.93.4-grml02) unstable; urgency=low
* New release based on upstream's git tree.
diff -Naurp initramfs-tools-0.93.4-grml02/mkinitramfs
initramfs-tools-0.93.4-grml02~hbpo02/mkinitramfs
--- initramfs-tools-0.93.4-grml02/mkinitramfs2010-08-16
23:03:01.0 +0200
+++ initramfs-tools-0.93.4-grml02~hbpo02/mkinitramfs2010-08-16
23:03:33.0 +0200
@@ -237,7 +237,7 @@ copy_exec /sbin/modprobe /sbin
copy_exec /sbin/depmod /sbin
copy_exec /sbin/rmmod /sbin
mkdir -p ${DESTDIR}/etc/modprobe.d
-cp -a /etc/modprobe.d/*.conf ${DESTDIR}/etc/modprobe.d/
+cp -a /etc/modprobe.d/* ${DESTDIR}/etc/modprobe.d/
# workaround: libgcc always needed on old-abi arm
if [ $DPKG_ARCH = arm ] || [ $DPKG_ARCH = armeb ]; then
- recompile fai 3.3.5 (current) for my local repo
Else all the machines get initrams-tools=0.97 installed that is shipped with
deb http://www.informatik.uni-koeln.de/fai/download lenny koeln
in /etc/apt/sources.list.
Furthermore I needed to add an extra functionality to live-initramfs:
Do dhcp in live-initramfs on the nic with the mac-address gotten from
pxelinux BOOTIF parameter (ipappend 2).
I will open an extra bug (wishlist) with a patch attached.
To make a long story short:
My Questions are:
Is this version I prepared/use (0.93.4-grml02~hbpo02) compatible with
lenny (for fai nfsroot !and! the installed machines)? - I get no error
messages when I use it (o:
Is every version of initramfs-tools after commit
553aa3742ca43b4ba4e87b2dea2c5d31cc43a124 - mkinitramfs: generate
pre-cached boot order file (and so also =0.94) incompatible with lenny
(breaks f.e. the plain lenny version of cryptsetup 1.0.6-7,... ?)?
mainly: Am I right or do I miss something? (o:
If so:
Would it make sense to downgrade initramfs-tools shipped with fai repo
of uni koeln?
Cheers
Holger Fischer
--
To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/4c69a86d.1020...@web.de
Bug#581596: linux-2.6: pxe-booting a qemu-kvm or kvm guest with virtio network (fai-client) produces kernel panic
Hallo, it's not a kernel problem. When installing the newer initramfs-tools from official lenny fai repo (http://www.informatik.uni-koeln.de/fai/download/lenny/, initramfs-tools_0.93.4-grml02_all.deb) pxe boots fine with virtio-net guests (both with plain lenny 2.6.26... and backported 2.6.32 from squeeze). Possibly you want to assign this bug initramfs-tools. Cheers Holger Fischer -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4c2a67dc.4060...@web.de
Bug#587580: linux-source-2.6.32: bonding (ifenslave) does not work with asix based usb-network-adapter (ax88178)
Package: linux-source-2.6.32
Severity: normal
Tags: patch
Hallo,
This patch from git.kernel.org fixes setting of mac address of asix usb-net
adapters,
MAC address setting is needed by ifenslave (mode active/backup).
Without this patch bonding seems to work with my ax88178 based,
but when making this device the active no packets are transmitted.
Switching back to the primary active device (e1000) works - no errors, oops,
panic.
When applying this patch to the current squeeze kernel sources (2.6.32-15),
compiling and installing it, the ax88178 based adapter works as expected in
bonding mode active/backup.
This works also on a lenny system with the backported squeeze kernel.
It would be nice if this patch could be included in squeeze.
P.S. Possibly this is related to bug 444043.
Thanks
Holger Fischer
diff --git a/drivers/net/usb/asix.c b/drivers/net/usb/asix.c
index 20e3460..9e05639 100644
--- a/drivers/net/usb/asix.c
+++ b/drivers/net/usb/asix.c
@@ -54,6 +54,7 @@ static const char driver_name [] = asix;
#define AX_CMD_WRITE_IPG0 0x12
#define AX_CMD_WRITE_IPG1 0x13
#define AX_CMD_READ_NODE_ID0x13
+#define AX_CMD_WRITE_NODE_ID 0x14
#define AX_CMD_WRITE_IPG2 0x14
#define AX_CMD_WRITE_MULTI_FILTER 0x16
#define AX88172_CMD_READ_NODE_ID 0x17
@@ -165,6 +166,7 @@ static const char driver_name [] = asix;
/* This structure cannot exceed sizeof(unsigned long [5]) AKA 20 bytes */
struct asix_data {
u8 multi_filter[AX_MCAST_FILTER_SIZE];
+ u8 mac_addr[ETH_ALEN];
u8 phymode;
u8 ledmode;
u8 eeprom_len;
@@ -732,6 +734,30 @@ static int asix_ioctl (struct net_device *net, struct
ifreq *rq, int cmd)
return generic_mii_ioctl(dev-mii, if_mii(rq), cmd, NULL);
}
+static int asix_set_mac_address(struct net_device *net, void *p)
+{
+ struct usbnet *dev = netdev_priv(net);
+ struct asix_data *data = (struct asix_data *)dev-data;
+ struct sockaddr *addr = p;
+
+ if (netif_running(net))
+ return -EBUSY;
+ if (!is_valid_ether_addr(addr-sa_data))
+ return -EADDRNOTAVAIL;
+
+ memcpy(net-dev_addr, addr-sa_data, ETH_ALEN);
+
+ /* We use the 20 byte dev-data
+* for our 6 byte mac buffer
+* to avoid allocating memory that
+* is tricky to free later */
+ memcpy(data-mac_addr, addr-sa_data, ETH_ALEN);
+ asix_write_cmd_async(dev, AX_CMD_WRITE_NODE_ID, 0, 0, ETH_ALEN,
+ data-mac_addr);
+
+ return 0;
+}
+
/* We need to override some ethtool_ops so we require our
own structure so we don't interfere with other usbnet
devices that may be connected at the same time. */
@@ -919,7 +945,7 @@ static const struct net_device_ops ax88772_netdev_ops = {
.ndo_start_xmit = usbnet_start_xmit,
.ndo_tx_timeout = usbnet_tx_timeout,
.ndo_change_mtu = usbnet_change_mtu,
- .ndo_set_mac_address= eth_mac_addr,
+ .ndo_set_mac_address= asix_set_mac_address,
.ndo_validate_addr = eth_validate_addr,
.ndo_do_ioctl = asix_ioctl,
.ndo_set_multicast_list = asix_set_multicast,
@@ -1213,7 +1239,7 @@ static const struct net_device_ops ax88178_netdev_ops = {
.ndo_stop = usbnet_stop,
.ndo_start_xmit = usbnet_start_xmit,
.ndo_tx_timeout = usbnet_tx_timeout,
- .ndo_set_mac_address= eth_mac_addr,
+ .ndo_set_mac_address= asix_set_mac_address,
.ndo_validate_addr = eth_validate_addr,
.ndo_set_multicast_list = asix_set_multicast,
.ndo_do_ioctl = asix_ioctl,
-- System Information:
Debian Release: squeeze/sid
APT prefers testing
APT policy: (500, 'testing')
Architecture: amd64 (x86_64)
Kernel: Linux 2.6.32-5-amd64 (SMP w/1 CPU core)
Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968)
Shell: /bin/sh linked to /bin/dash
--
To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive:
http://lists.debian.org/20100630003302.3736.66637.report...@hoogrzulisv001.intern.hoonet.org
Bug#581596: linux-2.6: pxe-booting a qemu-kvm or kvm guest with virtio network (fai-client) produces kernel panic
Package: linux-2.6 Version: 2.6.32-12 Severity: normal -- System Information: Debian Release: 5.0.4 APT prefers stable APT policy: (990, 'stable') Architecture: amd64 (x86_64) Kernel: Linux 2.6.26-2-amd64 (SMP w/4 CPU cores) Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/bash This bug was introduced by 2.6.32-12 and working fine with 2.6.32-11, I believe i firstly tested it with a 2.6.30 from squeeze. pxe-booting from a virtio device is not working with standard lenny kernel (2.6.26...) - also a kernel panic, that's why I backported a newer one. Some further information: qemu-kvm host: squeeze, uname -a:Linux hpedebsv03 2.6.32-3-amd64 #1 SMP Wed Feb 24 18:07:42 UTC 2010 x86_64 GNU/Linux dpkg -l |grep virt ii gvfs1.6.0-1+b1 userspace virtual filesystem - server ii libvirt-bin 0.8.0-2 the programs for the libvirt library ii libvirt00.8.0-2 library for interfacing with different virtu ii python-libvirt 0.8.0-2 libvirt Python bindings ii qemu-kvm0.12.3+dfsg-4 Full virtualization on x86 hardware ii virt-manager0.8.4-2 desktop application for managing virtual mac ii virt-viewer 0.2.1-1 Displaying the graphical console of a virtua ii virtinst0.500.3-1 Programs to create and clone virtual machine fai-server: lenny fai server installed with fai, nfsroot installed with a backported kernel, kernel compiled from original linux-2.6_2.6.32-12.dsc,etc... from sid on a fresh and clean fai lenny minimal installation. Best Regards Holger Fischer -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20100514055816.6604.70300.report...@hpedebsv19.hpe.hoonet.org
