Jeff Garzik wrote:
Sage Weil wrote:
Hi Jeff,
Do you still plan to replace bdb (and it's replication) with a
something based on paxos? I'm considering replacing the Ceph monitors
(which currently implement paxos, but in a very ceph-specific way)
with cld if it can meet the basic
Sage Weil wrote:
On Fri, 31 Jul 2009, Jeff Garzik wrote:
Sage Weil wrote:
Hi Jeff,
Do you still plan to replace bdb (and it's replication) with a something
based on paxos? I'm considering replacing the Ceph monitors (which
currently implement paxos, but in a very ceph-specific way) with cld
Just added per-operation authentication checks to chunkd. If the
authentication user does not own an object, they cannot GET or DELETE
it. LIST only shows objects owned by the authenticated user (making the
Owner tag a bit redundant).
The newly added 'auth' test verified that this works,
The URL for the release tarballs of Project Hail,
http://www.kernel.org/pub/software/network/distsrv/
was getting a little crowded, with all the tarballs in that one
directory. The tarballs have been moved into per-project sub-directories.
cld tarballs will live in
Pete Zaitcev wrote:
It's convenient to run cldcli without arguments if your DNS is set up.
Signed-off-by: Pete Zaitcev zait...@redhat.com
applied
--
To unsubscribe from this list: send the line unsubscribe hail-devel in
the body of a message to majord...@vger.kernel.org
More majordomo info
Pete Zaitcev wrote:
This seems copy-pasted from a daemon like Chunk that has a comment
to the tune of now that we have arguments parsed we can switch
to syslog. But in cldcli it's meaningless.
Signed-off-by: Pete Zaitcev zait...@redhat.com
applied
--
To unsubscribe from this list: send the
Pete Zaitcev wrote:
Let's change build ports to xxx82 to deconflict with tabled.
Signed-Off-By: Pete Zaitcev zait...@redhat.com
diff --git a/test/server-test.cfg b/test/server-test.cfg
index 38fcc43..02c39d4 100644
--- a/test/server-test.cfg
+++ b/test/server-test.cfg
@@ -5,11 +5,11 @@
/SSL
Pete Zaitcev wrote:
On Tue, 04 Aug 2009 21:00:36 -0400, Jeff Garzik j...@garzik.org wrote:
Speaking of chunkd...
I don't think I will have time this week to fix chunkd's failure on big
endian machines. If you (or anyone else listening...) have an
opportunity to look into that, that would
Pete Zaitcev wrote:
On Tue, 04 Aug 2009 21:40:22 -0400, Jeff Garzik j...@garzik.org wrote:
Another avenue is to sign up for a Fedora Account
(http://fedoraproject.org/wiki/Join) and get access to scratch builds
via koji (https://fedoraproject.org/wiki/Using_the_Koji_build_system).
Chunk
Below is my current work-in-progress, finishing the packet / message
logical separation. This will permit messages larger than a single
packet, which, in turn, permits GET and PUT messages to be more easily
handled as a single, contiguous message. DATA messages are eliminated.
This is a major
Attached is the current CLD replication patch, which takes CLD from
being a single-node service to a fully replicated, highly available
service.
The server implementation should be complete.
The current merge blocker is needed code in libcldc, which does not yet
properly hunt for a master,
Pete Zaitcev wrote:
This patch transitions from printf-like to syslog-like API for
libcldc diagnostics. In the future, this allows to drop app_log()
from chunkd (tabled didn't even grow it).
There's a small complication: one stdarg function cannot call
another, so we cannot just insert some
Pete Zaitcev wrote:
Add our listening information into the parameter file in CLD.
For now we only do that, and do not try to reorganize the CLD
namespace that Chunk uses, use full path as a parameter, and other
planned things. We do switch to asprintf though, it makes the
code more robust (no
Pete Zaitcev wrote:
Without this patch, open errors cannot be delivered to application
and opens get stuck forever. Also, who added -EINVAL again?! Stop
doing that, the user wants to know what failed: syscall or CLD protocol.
Signed-Off-By: Pete Zaitcev zait...@redhat.com
applied
--
To
Fabian Deutsch wrote:
Hey.
Reading about the release of XtremFS 1.0 [1] i stumbled accross
babudb [2]. babudb provides some log based format, requested in the hail
wiki [3] for CLD.
So I just wanted to let you know :)
- fabian
[1] http://www.xtreemfs.org/
Both are interesting.. I note
On 08/16/2009 09:10 AM, Pete Zaitcev wrote:
This patch probably has all that's necessary to flip the configuration
to use CLD to configure available Chunk nodes and dispose with
StorageNode forever. However, it did not complete all tests that's
necessary to declare victory. So for now we
On 08/25/2009 09:13 PM, John Quigley wrote:
Hey Folks:
I've just recently discovered the existence of this project, and am
excited by its prospects. I hopefully have relevant experience to help
with the effort, as I've spent the last four years working on a network
distributed storage
[slight change, error messages kept in array -jg]
Signed-off-by: Jeff Garzik jgar...@redhat.com
diff --git a/include/cld_msg.h b/include/cld_msg.h
index 01bda16..e4c8f28 100644
--- a/include/cld_msg.h
+++ b/include/cld_msg.h
@@ -257,5 +257,6 @@ struct cld_msg_event {
extern unsigned long long
On 08/28/2009 10:41 PM, Pete Zaitcev wrote:
With this patch, daemons that tabled needs may come up at any time
relatively to each other and tabled should be able to initialize itself.
To this end, we remove the extra sleep command, now unnecessary.
Unfortunately, one sleep is still needed for
On 09/30/2009 09:42 PM, Pete Zaitcev wrote:
- Add the fresh option ForceHost, for the sake of Fedora builds.
- Change the NID just to make it more interesting (this helps to
make sure we're not using 1 by accident).
Signed-off-by: Pete Zaitcevzait...@redhat.com
applied 1-2
--
To
On 10/13/2009 01:27 PM, Zack Perry wrote:
Hi,
While trying to find a less expensive way to validate my application targeting
AWS S3, I stumbled across Project Hail.
After reading the wiki, I downloaded cld, chunkd, and tabled tar balls, built
them on my Ubuntu 9.04 based VPS successfully.
(yeah, the summary line needs s/cld:/chunkd:/)
Just committed the following change to chunkd's storage backend API,
preparing chunkd for variable-length opaque binary keys.
commit 0a8c3225355de86484eba71279c646f7790d5bb1
Author: Jeff Garzik j...@garzik.org
Date: Wed Nov 4 00:44:55 2009
On 11/04/2009 12:51 AM, Jeff Garzik wrote:
(yeah, the summary line needs s/cld:/chunkd:/)
Just committed the following change to chunkd's storage backend API,
preparing chunkd for variable-length opaque binary keys.
commit 0a8c3225355de86484eba71279c646f7790d5bb1
Author: Jeff Garzikj
A command line client for chunkd, chcli, was just committed. It offers
the ability to GET (dumps to stdout) and PUT (input from cmd line arg).
Also, some quick searching revealed a command line tool for S3 (and thus
tabled), s3cmd: http://s3tools.org/s3cmd
Jeff
--
To
You wrote this insightful and pointed comment on IRC...
Comparing with every k/v service out there assumes that you're
growing a generic key/value service out of Chunk. You're essentially
admitting it openly.
This is an excellent point to raise. So let the begin at the
beginning, cover the
On 11/14/2009 01:30 AM, Pete Zaitcev wrote:
Signed-off-by: Pete Zaitcevzait...@redhat.com
applied 1-2
--
To unsubscribe from this list: send the line unsubscribe hail-devel in
the body of a message to majord...@vger.kernel.org
More majordomo info at
On 11/14/2009 01:32 AM, Pete Zaitcev wrote:
I got into habit of watching logs and there was too much useless
messaging in them. Some messages were stubs and reminders, so I left
them in the code, but commented out.
Signed-off-by: Pete Zaitcevzait...@redhat.com
I applied this, just to keep
On 11/14/2009 01:33 AM, Pete Zaitcev wrote:
This bug leads to object's data corrupt whenever the write is partial.
Signed-off-by: Pete Zaitcevzait...@redhat.com
applied
--
To unsubscribe from this list: send the line unsubscribe hail-devel in
the body of a message to
On 11/14/2009 01:36 AM, Pete Zaitcev wrote:
This patch adds what amounts to a background process that maintains
redundancy for object data. It is far from the complete solution.
For one thing, it does not verify checksums. But it's a start.
There's no way to turn this off, by intention. The
This is a simple test to determine memory usage for storing object
locations in RAM, indexed by node id, and optimized for
- rapid object id insertion, deletion
- easy method to determine all objects in a single node
Compile with
gcc -O -Wall `pkg-config glib-2.0
|8 +-
server/server.c | 287 ++---
test/pid-exists | 24 +++--
test/prep-db | 19 ++--
test/start-daemon | 24 -
test/stop-daemon | 18 ++--
9 files changed, 419 insertions(+), 59 deletions(-)
Jeff Garzik (1):
cld
On 11/18/2009 12:00 PM, Pete Zaitcev wrote:
On Wed, 18 Nov 2009 03:45:40 -0500, Jeff Garzikj...@garzik.org wrote:
I just updated the current CLD replication code to the latest upstream
cld.git code. It is now living on the replica branch of
git://git.kernel.org/pub/scm/daemon/cld/cld.git
On 11/26/2009 09:39 PM, Pete Zaitcev wrote:
This patch adds what amounts to a background process that maintains
redundancy for object data. It is far from the complete solution.
For one thing, it does not verify checksums. But it's a start.
There's no way to turn this off, by intention. The
On 11/28/2009 02:07 AM, Pete Zaitcev wrote:
On Sat, 28 Nov 2009 00:09:09 -0500
Jeff Garzikj...@garzik.org wrote:
On 11/27/2009 10:52 PM, Pete Zaitcev wrote:
The patched tabled can build and work with the old libcldc too.
Part of the reason for that is the continued presence of
On 11/27/2009 10:52 PM, Pete Zaitcev wrote:
This patch makes it so that tabled stops relying on libcldc using libevent.
Thus, libcldc would be free to drop libevent dependency.
The patched tabled can build and work with the old libcldc too.
Signed-Off-By: Pete Zaitcevzait...@redhat.com
---
On 11/27/2009 06:20 PM, cmcc...@alumni.cmu.edu wrote:
--- a/lib/common.c
+++ b/lib/common.c
@@ -56,6 +56,37 @@ const char *cld_errstr(enum cle_err_codes ecode)
return cld_errlist[ecode];
}
+/** Read from a file descriptor, resuming after interruptions.
+ *
+ * @param fd The
On 11/27/2009 09:23 PM, Colin McCabe wrote:
Signed-off-by: Colin McCabecmcc...@alumni.cmu.edu
---
lib/common.c | 29 +++--
1 files changed, 15 insertions(+), 14 deletions(-)
diff --git a/lib/common.c b/lib/common.c
index 68f60f8..db20e2a 100644
--- a/lib/common.c
On 11/28/2009 05:37 AM, Jeff Garzik wrote:
1) 'buf' is no longer nul-terminated, which means strtol() has become a
buffer overrun.
Whoops, it seems I am incorrect on this, according to
http://library.gnome.org/devel/glib/stable/glib-File-Utilities.html#g-file-get-contents
Ignore that point
On 11/28/2009 07:34 AM, Colin McCabe wrote:
Signed-off-by: Colin McCabecmcc...@alumni.cmu.edu
---
lib/common.c | 27 +--
1 files changed, 13 insertions(+), 14 deletions(-)
applied, thanks for the revisions
--
To unsubscribe from this list: send the line
On 11/29/2009 01:35 AM, Pete Zaitcev wrote:
From: Jeff Garzikjgar...@redhat.com
Move timer_init and friends so that they can be used by test/*.
Signed-Off-By: Pete Zaitcevzait...@redhat.com
applied 1-2
--
To unsubscribe from this list: send the line unsubscribe hail-devel in
the body of a
On 11/29/2009 01:35 AM, Pete Zaitcev wrote:
From: Jeff Garzikjgar...@redhat.com
Move timer_init and friends so that they can be used by test/*.
Signed-Off-By: Pete Zaitcevzait...@redhat.com
hmmm... cld now segfaults reliably in koji:
On 11/29/2009 02:34 PM, Pete Zaitcev wrote:
On Sun, 29 Nov 2009 06:11:44 -0500 Jeff Garzikj...@garzik.org wrote:
hmmm... cld now segfaults reliably in koji:
http://koji.fedoraproject.org/koji/taskinfo?taskID=1836079
Curious. It works fine here, of course (make distcheck).
Did you try to
On 11/29/2009 08:17 PM, Pete Zaitcev wrote:
Fedora 13 comes with db4.8 and apparently the compat-db4.7 is bust.
Let us link with 4.8 as a workaround.
Signed-Off-By: Pete Zaitcevzait...@redhat.com
---
configure.ac |5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
Not sure how
On 12/01/2009 08:18 PM, Pete Zaitcev wrote:
On Sun, 29 Nov 2009 20:38:45 -0500
Jeff Garzikj...@garzik.org wrote:
Interesting... I recall the root cause clearly, now: /usr/include/db.h
always refers to the latest installed db4, even if compat-db{,45,46} is
installed. Our configure recipe
On 12/06/2009 07:39 PM, Colin McCabe wrote:
I'm wondering if there's a function that will take a buffer full of
XDR data and an xdrproc_t, and print out a human readable version of
the data.
It seems like an obvious debugging tool, but I can't find a reference
to anything like this in the XDR
On 12/14/2009 02:23 AM, Colin McCabe wrote:
Hi all,
I think I'm going to respin this patch to be a lot smaller. Just
refactor one or two things, but leave the existing verbose and
applog system in place. Let me know what you think-- do you think I
should push ahead with HAIL_DEBUG etc.,
On 12/14/2009 07:32 PM, Colin McCabe wrote:
Signed-off-by: Colin McCabecmcc...@alumni.cmu.edu
applied, after fixing two problems:
* did not survive 'make distcheck'. common.h was needed in
include/Makefile.am.
* introduced build warnings in the server. added #include in server/cld.h.
On 12/14/2009 07:32 PM, Colin McCabe wrote:
Signed-off-by: Colin McCabecmcc...@alumni.cmu.edu
applied, after fixing two minor problems:
* added to include/Makefile.am
* #include'd stdbool.h
--
To unsubscribe from this list: send the line unsubscribe hail-devel in
the body of a message to
On 12/14/2009 07:32 PM, Colin McCabe wrote:
Signed-off-by: Colin McCabecmcc...@alumni.cmu.edu
---
include/cldc.h |5 +--
lib/cldc.c | 98 +++-
tools/cldcli.c |2 +-
3 files changed, 43 insertions(+), 62 deletions(-)
applied
On 12/14/2009 07:32 PM, Colin McCabe wrote:
Signed-off-by: Colin McCabecmcc...@alumni.cmu.edu
applied, with minor changes:
* include hail_log.h from server/cld.h, the header file common to all
server code. remove individual #include's in each .c file.
* fixed up #include of
On 12/14/2009 07:32 PM, Colin McCabe wrote:
Signed-off-by: Colin McCabecmcc...@alumni.cmu.edu
---
include/cldc.h |7 ++---
lib/cldc-dns.c | 80 +---
tools/cldcli.c | 12 +---
3 files changed, 41 insertions(+), 58 deletions(-)
On 12/14/2009 07:55 PM, Colin McCabe wrote:
I almost forgot-- Makefile.am needs to change.
Signed-off-by: Colin McCabecmcc...@alumni.cmu.edu
---
include/Makefile.am |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/include/Makefile.am b/include/Makefile.am
index
On 12/14/2009 07:32 PM, Colin McCabe wrote:
Signed-off-by: Colin McCabecmcc...@alumni.cmu.edu
---
server/cldu.c | 14 --
server/tabled.h |1 +
2 files changed, 13 insertions(+), 2 deletions(-)
ditto last message, for tabled
--
To unsubscribe from this list: send the
On 12/16/2009 12:10 AM, Pete Zaitcev wrote:
This flag was is a useless duplication and a fiction prompted by
getaddrinfo(). In reality it's not useful even if we decide to run
CLD over DCCP.
Signed-off-by: Pete Zaitcevzait...@redhat.com
---
server/storage.c |8
server/tabled.h
zaitcev and I typically maintain recent versions of cld/chunkd/tabled in
Fedora rawhide. This can be a useful source of SRPMs for Fedora
10/11/12, if people find those useful. Usually it takes a few days for
the built [S]RPMs to trickle out to the Fedora mirrors. Since we just
committed
Looking at the 'python-boto' package in Fedora, I see it ships with
boto's basic test suite. Notably, it ships with a test module that
exercises boto's S3 library routines.
Is there anyone that would be interested in copying (or directly use)
On 12/18/2009 12:09 PM, Pete Zaitcev wrote:
Since I have no clue about Python, I had trouble making their standard
testing harness to work, so I just copied the test into a local file,
and appended an invocation like this:
from boto.s3.connection import S3Connection
from boto.s3.connection
On 12/18/2009 12:16 PM, Jeff Darcy wrote:
On 12/18/2009 12:09 PM, Pete Zaitcev wrote:
1. I fixed the calling format issue long time ago. Mind that you have
to add a wildcard for it to work:
; Default port is 8081, we override it with -p 4499 in /etc/sysconfig/cld
_cld._udp IN SRV
On 12/20/2009 09:45 PM, Pete Zaitcev wrote:
Existing code writes checksums of something other than the object data.
Fix by summing the object data.
Signed-off-by: Pete Zaitcevzait...@redhat.com
---
server/object.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
This seems too
On 12/22/2009 04:41 PM, Pete Zaitcev wrote:
I'm looking into adding self-checking to chunkd. This involves basically
a process that re-reads everything stored in the chunkserver and verifies
that it's still ok. Nothing can be simpler, right?
So, current problems for which I'd like input are:
On 12/22/2009 08:40 PM, Pete Zaitcev wrote:
On Tue, 22 Dec 2009 17:43:58 -0500
Jeff Garzikj...@garzik.org wrote:
It is normal and reasonable to maintain global information about all
in-progress operations. Caching systems do that, for example, to ensure
multiple cache requests for object A
Working on libcldc's functionality requires moving all cldc_host
searching and manipulation out from chunkd/tabled and into libcldc
proper. Part of this involves changing the integration points related
to SRV record lookups, so as to accomodate input from a file during
testing (thus avoiding the
On 12/26/2009 01:34 AM, Pete Zaitcev wrote:
diff --git a/server/be-fs.c b/server/be-fs.c
index 89635f8..6551c83 100644
--- a/server/be-fs.c
+++ b/server/be-fs.c
@@ -75,7 +75,7 @@ int fs_open(void)
}
if (!tchdbsetmutex(hdb))
- goto out_hdb;
+ goto
On 12/26/2009 01:34 AM, Pete Zaitcev wrote:
Signed-off-by: Pete Zaitcevzait...@redhat.com
---
server/chunkd.h |1 -
1 file changed, 1 deletion(-)
commit 6fe147ad2f8833e162d92277ec6827520c5497ba
Author: Masterzait...@lembas.zaitcev.lan
Date: Fri Dec 25 23:18:50 2009 -0700
Unused
On 12/27/2009 06:58 PM, Pete Zaitcev wrote:
This way we create a set of methods that can be used by self-check
to list existing objects.
Signed-off-by: Pete Zaitcevzait...@redhat.com
applied 1-2 quickly, since those were easy to review and apply.
I need more time to think about and review
On 01/03/2010 07:45 PM, Colin McCabe wrote:
Another attempt at escaping the overly-aggressive spam filters...
GCC has a nifty ability to typecheck varargs functions that follow the printf
conventions. This change uses that in HAIL_LOG. If you build with -Wall, gcc
can now give warnings about
On 01/05/2010 02:27 AM, Pete Zaitcev wrote:
Existing tests only excercised operations with relatively small objects.
It did not test pipelining of object data in sufficient degree. So, let's
have a better test case for this (large-object.c).
We also change the existing basic-object.c to match.
On 01/05/2010 11:02 PM, Pete Zaitcev wrote:
On Tue, 05 Jan 2010 04:13:36 -0500
Jeff Garzikj...@garzik.org wrote:
applied... chunkd needs these fixes also, yes?
No, it appears correct as-is. Although, the large-object test in
the suite does not check the data it transmits, so if chunkd were
On 01/10/2010 10:22 AM, Colin McCabe wrote:
Signed-off-by: Colin McCabecmcc...@alumni.cmu.edu
A very similar patch is needed for chunkd's server/cldu.c... did my
spam filter eat it?
Jeff
--
To unsubscribe from this list: send the line unsubscribe hail-devel in
the body of a
On 01/10/2010 10:00 AM, Colin McCabe wrote:
This patch moves CLD from using manual data serialization to using XDR (via
rpcgen). Both the packet header and the message body are now serialized and
deserialized using XDR. This makes it easy to have a variable-length packet
header, as well as a
On 01/11/2010 07:25 AM, Akinobu Mita wrote:
The volume directory may have 2^16 (=65536) directories because
prefix length of object pathname is 4 bytes. But ext3 filesystem can
only have EXT_LINK_MAX (32000) directories.
This changes the prefix length from 4 bytes to 3 bytes and breaks
the
On 12/27/2009 06:59 PM, Pete Zaitcev wrote:
This patch adds a background process that periodically verifies the
integrity of all known objects. Objects that are found found faulty
are made invisible to applications. This way, only known good objects
remain visible and applications may implement
On 01/11/2010 07:25 AM, Akinobu Mita wrote:
- Define _FILE_OFFSET_BITS
- Use uint64_t to hold object length and buffer length
- Don't pass too much buffer to writev. writev returns -1 if the sum of
the iov_len values overflows an ssize_t value
Signed-off-by: Akinobu
On 01/12/2010 01:25 PM, Colin McCabe wrote:
util.c: In function ‘objid_init’:
util.c:331: warning: integer constant is too large for ‘long’ type
with
[cmcc...@stargazer tabled]$ gcc --version
gcc (GCC) 4.4.1 20090725 (Red Hat 4.4.1-2)
Now that I wrote a little test program, I can see that gcc
On 01/12/2010 10:50 PM, Akinobu Mita wrote:
This patch makes sense, but it also raises the question of whether or not
we
should move to a two-level directory scheme, eg.
123/456/7890ABCDEF
rather than
123/4567890ABCDEF
to limit the size of the top-level directories. It
On 01/10/2010 10:00 AM, Colin McCabe wrote:
This patch moves CLD from using manual data serialization to using XDR (via
rpcgen). Both the packet header and the message body are now serialized and
deserialized using XDR. This makes it easy to have a variable-length packet
header, as well as a
My general comment on the libcldc API is that it is a low-level API, and
definitely not very user friendly at all. ncld or other APIs would
improve life quite a bit. Making life easier for programmers, and
cutting out tons of code, is a plus :)
Jeff, just returned from camping
On 01/14/2010 11:03 PM, Pete Zaitcev wrote:
The lib/ is a vestige of days when we stored files in a local filesystem,
not Chunk. Just kill it.
And as for CLD, it's a last resort configuration, not something we should
encourage. So comment it out.
Signed-Off-By: Pete Zaitcevzait...@redhat.com
On 01/14/2010 11:11 PM, Pete Zaitcev wrote:
Make the cell configurable. Among other things we need this for is
that you cannot migrate a tabled over when DB format changes, short
of running 2 CLDs.
We also add a few thoughts to the documentation, prompted by re-reading
of the document while
On 01/14/2010 11:12 PM, Pete Zaitcev wrote:
I find it too tiresome to monitor the status with kill -USR1.
This is more fun. The patch seems not too ugly regarding the factoring
of the cli_xxx methods.
Signed-Off-By: Pete Zaitcevzait...@redhat.com
---
doc/etc.tabled.conf |3
On 01/14/2010 11:13 PM, Pete Zaitcev wrote:
This problem was with us for a while, and even with this fix our start-up
is not reliable. But at least we will not be 100% guaranteed to hang as
before when restarting too quickly. So although the whole area needs some
serious reworking, this specific
On 01/14/2010 11:13 PM, Pete Zaitcev wrote:
Every time CLD crashes, I find my logs filled up with error messages.
But running without them is no good either, so I don't want to remove
this message entirely. Let us try some filtering.
Actually, I am still considering if we should not reset the
On 02/03/2010 08:45 AM, Colin McCabe wrote:
When we create a static buffer for an inode name, and treat it like a
null-terminated string, it needs to be of length CLD_INODE_NAME_MAX + 1 so
that it can hold the NULL-terminator.
In cldc_del and cldc_open, we should check that the user-submitted
On 01/20/2010 05:56 PM, Pete Zaitcev wrote:
Is there a way to cancel an outstanding lock request? How? You seem
to think that there's no problem.
Actually I think an cmo_close on a handle that has outstanding
requests of any kind should drop them, so I was incorrect about
killing the session
On 02/03/2010 04:45 PM, Colin McCabe wrote:
On Tue, Feb 2, 2010 at 10:35 PM, Jeff Garzikj...@garzik.org wrote:
I will continue whittling down the patch until it just contains the XDR
changes themselves.
In tools/Makefile.am, I don't think you need $(top_srcdir)/lib any
more, since
On 02/07/2010 02:00 AM, Pete Zaitcev wrote:
Hi, Jeff Colin:
It looks like you broke something in CLD, not sure if server or client.
There are two possibly related bugs. But first, here's the messages
(The chunkd is run with -D). Note that I have 2 servers listed in DNS
(both on port 4499), but
On 02/07/2010 02:15 PM, Pete Zaitcev wrote:
The traditional CLD API is too difficult to program. In particular,
switching from a rigid Group policy to arbitrary paths in existing
clients in Chunk and tabled turned out to be next to impossible.
The issue is due to the fundamental nature of the
On 02/08/2010 08:46 PM, Pete Zaitcev wrote:
On Mon, 08 Feb 2010 07:32:56 -0500
Jeff Garzikj...@garzik.org wrote:
- log param in cldc-dns should not be made conditional as an
undocumented side effect of ncld
I pulled, and it appears that you merged a code that crashes at the
first error
On 02/08/2010 08:49 PM, Pete Zaitcev wrote:
When I run autogen.sh, this message is printed:
lib/Makefile.am:13: `%'-style pattern rules are a GNU make extension
It's because of this:
%_rpc.h: %_rpc.x
rpcgen -h $$@
Is the above a concern?
It is likely a concern, yes. For
On 02/08/2010 09:17 PM, Pete Zaitcev wrote:
On Mon, 08 Feb 2010 21:06:17 -0500
Jeff Garzikj...@garzik.org wrote:
- if (cldc_getaddr(host_list, hostb, NULL))
+ if (cldc_getaddr(host_list, hostb,ncld_log))
return 1001;
Logging pointer should be supplied by the
On 02/08/2010 11:42 PM, Pete Zaitcev wrote:
No new function just yet, only a switch-over.
Signed-Off-By: Pete Zaitcevzait...@redhat.com
---
server/cldu.c | 789 +---
1 file changed, 215 insertions(+), 574 deletions(-)
Looks good, but needs a
On 02/09/2010 05:34 AM, Jeff Garzik wrote:
On 02/07/2010 02:00 AM, Pete Zaitcev wrote:
Hi, Jeff Colin:
It looks like you broke something in CLD, not sure if server or client.
There are two possibly related bugs. But first, here's the messages
(The chunkd is run with -D). Note that I have 2
;
exit(0);
commit 1c7d8de3e44e641651f766d6863780343ecaa5a8
Author: Jeff Garzik j...@garzik.org
Date: Wed Feb 17 20:14:04 2010 -0500
Cosmetic renaming related to ncld API.
s/fhp/fh/
s/ses/sess/
s/nsp/nsess/
And use a more standard way of calling callbacks
commit a58d541cf4357ab0aa7ae5e21765de35f65c6177
Author: Jeff Garzik j...@garzik.org
Date: Sat Mar 6 17:44:51 2010 -0500
tabled: fix key corruption due to string overrun
When obtaining the path (and thus, the object key) from the URI, it went
unnoticed that the URI-based
commit cb0ed2014b5aa68e6d0a2ffdad031b2d9ee31159
Author: Jeff Garzik j...@garzik.org
Date: Sun Mar 7 07:56:55 2010 -0500
[test] start-daemon: disable tabled debug output by default
Signed-off-by: Jeff Garzik jgar...@redhat.com
diff --git a/test/start-daemon b/test/start-daemon
On 03/15/2010 10:02 PM, Akinobu Mita wrote:
If the request key is not null-terminated and the debug output is enabled,
it prints invalid and unsafe value. Add a precision to fix it.
Signed-off-by: Akinobu Mitaakinobu.m...@gmail.com
---
server/server.c |3 ++-
1 files changed, 2
On 03/16/2010 05:59 AM, Akinobu Mita wrote:
At the beginning of stc_parse_key(), st_object is allocated twice for
the same variable.
Signed-off-by: Akinobu Mitaakinobu.m...@gmail.com
---
lib/chunkdc.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
good catch, applied
--
To
On 03/28/2010 09:57 PM, Pete Zaitcev wrote:
Signed-off-by: Pete Zaitcevzait...@redhat.com
---
server/storage.c |3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
applied
--
To unsubscribe from this list: send the line unsubscribe hail-devel in
the body of a message to
On 04/01/2010 07:01 AM, Samba - BoYang wrote:
hi, *
CLD doesn't build on db-4.3 on suse 11, since db-4.3 uses deprecated
structure members DBC-c_xxx(c_close(), etc) instead of DBC-xxx. :-)
It won't build on db-4.4, either. probably won't build on db-4.5, as
db-5.0 says DBC-xxx was
On 03/31/2010 08:43 PM, Pete Zaitcev wrote:
diff --git a/server/server.c b/server/server.c
index 3208e0f..2d68ee6 100644
--- a/server/server.c
+++ b/server/server.c
@@ -55,7 +55,7 @@ static struct argp_option options[] = {
Store database environment in DIRECTORY. Default:
1 - 100 of 196 matches
Mail list logo