[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [libvirt] [PATCH] support compressed crashdump of guests



On Thu, Oct 21, 2010 at 12:02:11PM +0900, KAMEZAWA Hiroyuki wrote:
> 
> Now, virsh dump doesn't support compresses dump.
> This patch adds GZIP and LZOP option to virsh dump and support
> it at qemu coredump. (AFAIK, LZOP is available on RHEL6.)
> 
> When I did 4G guest dump,
> (Raw)	3844669750 
> (Gzip)	1029846577
> (LZOP)	1416263880 (faster than gzip in general)
> 
> This will be a help for a host where crash-dump is used
> and several guests works on it.
> 
> help message is modified as this.
>   NAME
>     dump - dump the core of a domain to a file for analysis
> 
>   SYNOPSIS
>     dump [--live] [--crash] [--gzip] [--lzop] <domain> <file>
> 
>   DESCRIPTION
>     Core dump a domain.
> 
>   OPTIONS
>     --live           perform a live core dump if supported
>     --crash          crash the domain after core dump
>     --gzip           gzip dump(only one compression allowed
>     --lzop           lzop dump(only one compression allowed
>     [--domain] <string>  domain name, id or uuid
>     [--file] <string>  where to dump the core
> 
> Tested on Fedora-13+x86-64.
> 
> Note: for better compression, we may have to skip pages filled by
> zero or freed pages. But it seems it's qemu's works.

  The patch looks relatively simple and clean, I still have one comemnt
below though. It also seems to me that any use of the dump need a follow
up decompression before it can be fed to debug tools (gdb ...) as I
don't think they handle compressed dumps natively.

  The other comment on principle about this patch is that for saving
compression we use a qemu.conf option, maybe we should instead use
a new option there for core compression, or simply reuse the same
option for both (core being just a special kind of save). You will
also note that save_image_format takes more options than lzop or gzip.
If doing so one doesn't need to change virsh too.

  It's a trade off. If using the configuration option you need to plan
in advance the fact you will be dumping compressed core, if you expose
it at the API level itself, you can activate it anytime, and that may be
more useful for example when interacting with a customer facing an
issue needing debug. So both ways are acceptable to me.

> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa hiroyu jp fujitsu com>
> ---
>  include/libvirt/libvirt.h.in |    2 ++
>  src/qemu/qemu_driver.c       |   23 +++++++++++++++++++----
>  tools/virsh.c                |   10 +++++++++-
>  3 files changed, 30 insertions(+), 5 deletions(-)
> 
> Index: libvirt-0.8.4/src/qemu/qemu_driver.c
> ===================================================================
> --- libvirt-0.8.4.orig/src/qemu/qemu_driver.c
> +++ libvirt-0.8.4/src/qemu/qemu_driver.c
> @@ -5710,7 +5710,7 @@ cleanup:
>  
>  static int qemudDomainCoreDump(virDomainPtr dom,
>                                 const char *path,
> -                               int flags ATTRIBUTE_UNUSED) {
> +                               int flags) {
>      struct qemud_driver *driver = dom->conn->privateData;
>      virDomainObjPtr vm;
>      int resume = 0, paused = 0;
> @@ -5720,6 +5720,14 @@ static int qemudDomainCoreDump(virDomain
>          "cat",
>          NULL,
>      };
> +    const char *zargs[] = {
> +	"gzip",
> +	NULL,
> +    };
> +    const char *lzargs[] = {
> +	"lzop",
> +	NULL,
> +    };
>      qemuDomainObjPrivatePtr priv;
>  
>      qemuDriverLock(driver);
> @@ -5787,9 +5795,16 @@ static int qemudDomainCoreDump(virDomain
>      }
>  
>      qemuDomainObjEnterMonitorWithDriver(driver, vm);
> -    ret = qemuMonitorMigrateToFile(priv->mon,
> -                                   QEMU_MONITOR_MIGRATE_BACKGROUND,
> -                                   args, path, 0);
> +    if (flags & VIR_DUMP_GZIP)
> +	ret = qemuMonitorMigrateToFile(priv->mon,
> +		QEMU_MONITOR_MIGRATE_BACKGROUND, zargs, path, 0);
> +    else if (flags & VIR_DUMP_LZOP)
> +	ret = qemuMonitorMigrateToFile(priv->mon,
> +		QEMU_MONITOR_MIGRATE_BACKGROUND, lzargs, path, 0);
> +    else
> +	ret = qemuMonitorMigrateToFile(priv->mon,
> +		QEMU_MONITOR_MIGRATE_BACKGROUND, args, path, 0);
> +

  I'm wondering what happens if the compression command is not present,
if the current error message we get back from qemu is clear enough then
fine, but otherwise we may have to check for the presence of the
compressor program.

>      qemuDomainObjExitMonitorWithDriver(driver, vm);
>      if (ret < 0)
>          goto endjob;
> Index: libvirt-0.8.4/tools/virsh.c
> ===================================================================
> --- libvirt-0.8.4.orig/tools/virsh.c
> +++ libvirt-0.8.4/tools/virsh.c
> @@ -1751,6 +1751,8 @@ static const vshCmdInfo info_dump[] = {
>  static const vshCmdOptDef opts_dump[] = {
>      {"live", VSH_OT_BOOL, 0, N_("perform a live core dump if supported")},
>      {"crash", VSH_OT_BOOL, 0, N_("crash the domain after core dump")},
> +    {"gzip", VSH_OT_BOOL, 0, N_("gzip dump(only one compression allowed")},
> +    {"lzop", VSH_OT_BOOL, 0, N_("lzop dump(only one compression allowed")},

  what about bzip2 and xz, the first one is probably way too slow, but
  we accept it for saves.

>      {"domain", VSH_OT_DATA, VSH_OFLAG_REQ, N_("domain name, id or uuid")},
>      {"file", VSH_OT_DATA, VSH_OFLAG_REQ, N_("where to dump the core")},
>      {NULL, 0, 0, NULL}
> @@ -1778,7 +1780,13 @@ cmdDump(vshControl *ctl, const vshCmd *c
>          flags |= VIR_DUMP_LIVE;
>      if (vshCommandOptBool (cmd, "crash"))
>          flags |= VIR_DUMP_CRASH;
> -
> +    if (vshCommandOptBool (cmd, "gzip"))
> +	flags |= VIR_DUMP_GZIP;
> +    if (vshCommandOptBool (cmd, "lzop"))
> +        flags |= VIR_DUMP_LZOP;
> +    if ((flags & (VIR_DUMP_GZIP | VIR_DUMP_LZOP))
> +	 == (VIR_DUMP_GZIP | VIR_DUMP_LZOP))
> +	return FALSE;

  I think an error message need to be provided stating that both options
are mutually exclusive.

  Otherwise that looks fine to me,

Daniel

-- 
Daniel Veillard      | libxml Gnome XML XSLT toolkit  http://xmlsoft.org/
daniel veillard com  | Rpmfind RPM search engine http://rpmfind.net/
http://veillard.com/ | virtualization library  http://libvirt.org/


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]