te operation performed by a Linux cache
manager that uses a disk cache.
For those who are interested, the original commit was
34ffc9cd7d7eed62229704ad0e1d327f076ea7b6, reviewed at
http://gerrit.openafs.org/903
Cheers,
Simon.
—
Simon Wilkinson
AuriStor Inc.
___
ver, when I originally wrote the
splice() code it had a significant performance benefit, so I would imagine
disabling it will have a speed impact.
I can confirm that the AuriStor client takes a different approach to fixing
this issue, and has no performance hit from doing so.
Cheers,
Simon.
etween ‘private’ and ‘public’ fileservers.
Keep more sensitive data on fileservers that aren’t exposed to the internet.
7) Consider keeping some of your database servers hidden from the public
internet
Cheers,
Simon
—
Simon Wilkinson
AuriStor Inc.
___
OpenA
On 5 Aug 2014, at 14:08, Brandon Allbery ballb...@sinenomine.net wrote:
On Tue, 2014-08-05 at 09:30 +0200, Alex wrote:
Now, I didn't find in the admin guide or wiki[1] some useful
information
about client's firewall, but I could find some information on the
Internet saying that client
On 5 Aug 2014, at 16:09, Brandon Allbery ballb...@sinenomine.net wrote:
On Tue, 2014-08-05 at 16:08 +0100, Simon Wilkinson wrote:
The complication is that firewalls/NATs only preserve these mappings
for a finite length of time. We attempt to keep them open through
regular fileserver pings
On 5 Aug 2014, at 15:57, Brandon Allbery ballb...@sinenomine.net wrote:
Not yet. This is one of the things rxgk is supposed to address; we can
then use any GSSAPI-provided service. (The Globus stuff included a
minimal GSSAPI support mechanism with a number of shortcomings, IIRC.)
The Globus
On 5 Aug 2014, at 17:21, Alex euergetiko...@gmail.com wrote:
yes, what I meant is that I need the client to be aware that some other
client is editing, (and refresh the cache), which is the function of
callback if I am not mistaken. As I understand, this is not possible
behind a NAT
On 24 Jan 2014, at 07:48, Harald Barth h...@kth.se wrote:
You are completely right if one must talk to that server. But I think
that AFS/RX sometimes hangs to loong on waiting for one server
instead of trying the next one. For example for questions that could
be answered by any VLDB. I'm
On 23 Jan 2014, at 11:37, mi...@task.gda.pl mi...@task.gda.pl wrote:
Recently I came across an issue where 'make install' which was installing
software to AFS was taking *much* more time on a Solaris client than on
Linux client. The issue turned out to be a lack of optimization on VFS layer
On 20 Dec 2013, at 19:11, nicolas prochazka prochazka.nico...@gmail.com wrote:
so discon mode cannot work ?
and is it possible to define callback expiration time ? ( hacking code is a
solution for me, 2h=24h)
I suspect, but haven't checked, that the problem is that we're expiring volume
On 20 Mar 2013, at 18:01, Jeffrey Altman wrote:
I suspect the correct solution is to add
#define _FILE_OFFSET_BITS64
to the appropriate src/config/param.*.h file for the platform in
question in the #ifndef UKERNEL section.
The correct fix to this is to use the autoconf macro that
On 14 Mar 2013, at 22:05, Måns Nilsson wrote:
# cd src/venus/
# LC_ALL=C make # not all people read swedish compiler errors fluently
/opt/SUNWspro/bin/cc -I/usr/heimdal/include-O
-I/tank/scratch/openafs-1.6.2/src/config
-I/tank/scratch/openafs-1.6.2/include -I. -I. -mt
On 12 Dec 2012, at 20:47, Brandon Allbery wrote:
On Wed, Dec 12, 2012 at 3:30 PM, Steve Gaarder gaar...@math.cornell.edu
wrote:
THanks. Do I need to restart the afs servers in order to have them use the
new key?
Shouldn't be.
There's a race in the key selection algorithms for
On 11 Dec 2012, at 21:29, Brandon Allbery wrote:
This is what SRV records are for, yes. Ideally, the CellServDB on clients is
for legacy use with old cells
Sadly, there are loads of situations where the client CellServDB is vital.
There are far too many SOHO broadband routers that come
Any ideas, or suggestions for how else to find out what might be going
on, would be much appreciated. I haven't found anything suggestive by
searching.
This sounds a like you are waiting for a callback break to complete, or to time
out. You should be able to verify this by looking at a
On 19 Nov 2012, at 07:15, Jakub Moscicki wrote:
Thanks for this analysis. The increased UDP works pretty well for us at CERN
so far - albeit one limit gone other limits appear more pronounced.
I'm interested in what other limits you are hitting. I'm very aware of the
problems with the
On 9 Oct 2012, at 10:24, Dan Van Der Ster wrote:
We currently run fileservers with udpsize=2MB, and at that size we have a 30
client limit in our test environment. With a buffer size=8MB (increased
kernel max with sysctl and fileserver option), we don't see any dropped UDP
packets during
On 7 Nov 2012, at 21:52, David Bear wrote:
while my kernel is 2.6.18-308.16.1.el5, attempts to yum install kmod-openafs
come back telling me that there is a missing dependancy on kernel-x86_64 =
2.6.18-194.26.1.el5 .. This makes no sense to me. At least this is how I
read the following:
On 26 Oct 2012, at 23:40, Jeffrey Altman wrote:
I have significant concerns about the design of TCP OOB as it was
described at EAKC2012.
On the contrary, I think Out of Band support for AFS is a very interesting
prospect. As I discussed with Andrew and Hartmut in Edinburgh, I think we could
On 25 Sep 2012, at 18:58, Russ Allbery wrote:
Steve Simmons s...@umich.edu writes:
The last bit of discussion on perlafs with openafs 1.6 seemed to end
with the note below. If there's been progress on this, can anyone point
us at patches/whatever? We've got a person here who's trying to
A dearth of release managers, for example. Among the other things that
dearth is a bigger problem than is the absence of nightly builds. :)
This is also the reason that 1.4.x has stagnated - it's been 16 months since
1.4.14.1, and getting on for two years since 1.4.14. Andrew has been doing
On 17 Sep 2012, at 19:54, Troy Benjegerdes wrote:
If 'rebuild with debug' symbols is the answer to find the segfault, then why
don't we change './regen ./configure make check' to turn on debug
symbols
by default (at least in master.. we can turn it back off in a release)
If you are
On 17 Sep 2012, at 17:25, David Boyes wrote:
'make check' on a single machine will never give you useful testing results
other than to find packaging or smoke test errors, which aren't all that
helpful overall.
I agree with you with regards to crowd sourced testing, but I just wanted to
On 17 Sep 2012, at 23:18, Andrew Deason wrote:
I don't think you can make that say something based on 1.6.1, since the
head of the 1.6.x branch right now is a different branch than 1.6.1. I
mean, if git-version said something like this is 1.6.1 plus N patches,
that would be incorrect.
On 4 Sep 2012, at 16:38, Marc Dionne wrote:
That's very useful. It looks like dput gets called with the global
lock held, and it can potentially call back into AFS code through
afs_dentry_iput.
Might have to look at the afs_syscall_pioctl in more detail (it's
heavily ifdef'ed and hard to
On 14 Jun 2012, at 11:45, Jayen Ashar ja...@science.unsw.edu.au wrote:
Is there any chance this has changed in the last 9 years?
The details of how cross-realm users are created hasn't changed,
However, I don't think this is applicable to your situation. What you should do
is list both your
On 10 Jun 2012, at 06:41, Måns Nilsson wrote:
can get tokens, and [I believe] I
am a superuser, but I seem to have very few permissions.
Is jayen.admin listed in the UserList file?
Firstly, the UserList doesn't affect the behaviour of the ptserver. It
determines the user's permissions
On 3 Jun 2012, at 06:18, Jayen Ashar wrote:
On Sat, Jun 2, 2012 at 11:07 PM, Simon Wilkinson
simonxwilkin...@gmail.com wrote:
On 2 Jun 2012, at 01:47, Jayen Ashar wrote:
Yes. This should work, provided you can set up a cross realm trust between
the active directory realm, and the one
On 2 Jun 2012, at 01:47, Jayen Ashar wrote:
Would setting up our own realm for the AFS server work? Could all
users would be authenticated cross-realm? (We are not concerned with
cross-realm attacks at the moment.) Would any changes be needed to
the users' KDCs?
Yes. This should work,
On 9 May 2012, at 11:47, Stefan Michael Guenther wrote:
May 9 12:30:36 DELUH01-LSV001 kernel: afs: Tokens for user of AFS id 1010
for cell in-put.de are discarded (rxkad error=19270407, server
192.168.224.113)
No tokens, no rights?
Yes.
That error is security object was passed a bad
On 23 Apr 2012, at 12:59, Stefan Michael Guenther wrote:
Hi,
our system is a Heimdal Kerberos together with an OpenAFS 1.6.0.1 on a Ubuntu
11.10.
My first attempt was to add myself to the group of administrators:
root@intranet:~# pts adduser stefan system:administrators -cell
On 19 Apr 2012, at 19:54, Derrick Brashear sha...@gmail.com wrote:
Use the AFSCONF environment variable (takes a path) to point to a directory
per dbserver, with a cellservDB in that directory listing *only* that
dbserver for the cell.
Sadly, AFSCONF is only used if a configuration file
On 7 Apr 2012, at 15:50, Jason Edgecombe wrote:
Gerrit commit http://gerrit.openafs.org/7108 includes the previously
discussed /vicepXX/NeverAttach flag. The commit has been merged into the
master branch.
I can't close ticket 130561. Would someone close it, please?
Done. We really
On 6 Apr 2012, at 17:43, Darren Patterson wrote:
Mock version:
rhel-afs-test6x:/root# rpm -q mock
mock-1.1.11-1.el6.rf.noarch
This looks like mock is evaluating the srpm outside of the chroot. Given that
we know that this works with the mock from EPEL (1.1.22) could you try running
your
On 6 Apr 2012, at 19:22, Darren Patterson wrote:
Are you building the kmods against the CentOS kernel or the RHEL kernel? I
realize they are binary compatible. I'm grasping at straws at this point
since the 1.6.0 srpm (and everything else I build) builds fine with my mock
setup.
There was a change made to the spec file at the last minute in the 1.6.1
release process after problems were encountered with the mock builders.
Unfortunately this change includes an unnecessary kernel-devel dependency.
Sadly, whilst the problem was pointed out during the review, an unmodified
On 30 Mar 2012, at 11:54, Thomas Smith wrote:
Can someone explain what this setting does please? Just wondering if it's
encrypting communications while acquiring tokens ('auth') or providing some
kind of integrity checks to help avoid or catch data corruption ('data
integrity') or
On 29 Mar 2012, at 20:05, Brandon Allbery wrote:
SRV records help somewhat with the dbserver issue,
Sadly SRV records don't help quite as much as you might like. There are still
far too many SOHO routers on the market that claim to have a forwarding DNS
server, but just drop SRV records on
On 23 Mar 2012, at 18:52, Eric Chris Garrison wrote:
I've searched for this, and it appears mmap operations used to be a
problem. At what version of OpenAFS (server and client) should we be at
for this to not be a problem?
There's no difference on the server between files that are written
On 22 Mar 2012, at 15:09, Andrew Deason wrote:
On Thu, 22 Mar 2012 10:18:44 -0400
chas williams - CONTRACTOR c...@cmf.nrl.navy.mil wrote:
but besides this limit, there is also another determining factor in
rx. rx, like tcp, negotiates a window of data to send before waiting
for an ack
On 22 Mar 2012, at 15:23, Andrew Deason wrote:
On Thu, 22 Mar 2012 15:20:20 +
Simon Wilkinson simonxwilkin...@gmail.com wrote:
That limit is imposed because it is the point at which the current RX
implementation loses the queue efficiency/throughput tradeoff. You can
run with a larger
On 21 Feb 2012, at 16:25, Jeff Blaine jbla...@kickflop.net wrote:
-bash-3.2# /sbin/insmod /usr/vice/etc/modload/libafs-2.6.18-308.el5.mp.ko
insmod: error inserting '/usr/vice/etc/modload/libafs-2.6.18-308.el5.mp.ko':
-1 Unknown symbol in module
You either need to insmod exportfs first, or
On 17 Feb 2012, at 15:53, Andrew Deason wrote:
On Fri, 17 Feb 2012 14:51:31 +0100
Natxo Asenjo natxo.ase...@gmail.com wrote:
Apparently the kmod-openafs package is not for the same version so it
does not install.
That's because there apparently are no kmod packages for your kernel
On 17 Feb 2012, at 17:45, Natxo Asenjo wrote:
error: Failed build dependencies:
kernel-devel-i386 = 2.6.32-220.4.2.el6 is needed by
openafs-1.6.0-1.el6.i386
You need to specific i686 as the target.
S.
___
OpenAFS-info mailing list
On 17 Feb 2012, at 18:21, Natxo Asenjo wrote:
checking for linux kernel module build works... no
configure: error: Fix problem or use --disable-kernel-module...
See `config.log' for more details.
+ exit 1
error: Bad exit status from /var/tmp/rpm-tmp.Ak15pA (%build)
As it says, see
On 10 Feb 2012, at 04:29, Jeffrey Altman wrote:
That field was commented out before OpenAFS and maxPacketSize was not
referenced anywhere in the source code. What restoring that field to
rx.h does is force the struct rxevent * values that follow it to become
properly aligned in 64-bit
On 12 Feb 2012, at 00:02, muayad y wrote:
when maxPacketSize has been uncommented, it did change rx_connection struct
behavoir, rx_connection is being called on vos (userspace code)
The question is why this change actually improves performance - we need to
understand that, so we can tell
On 11 Dec 2011, at 10:42, Lars Schimmer wrote:
The empty line ahead of salvager was a problem. Good to know those make
problems, but is it really needed to be as strict as it is now?
Originally, the BosConfig file wasn't supposed to be user editable - you were
meant to use the 'bos'
On 2 Dec 2011, at 18:01, Stephan Wiesand stephan.wies...@desy.de wrote:
Thanks a lot. I'll try to find them in the source and get rid of them. Still
hoping for a hint making this search more efficient, though.
Look in src/rx/rx.c From memory, I think you are looking for the
On 22 Nov 2011, at 20:28, Jeff Blaine wrote:
I'm a little confused. I just had to turn on
allow_weak_crypto in a RHEL6 kerberos client's
/etc/krb5.conf to be able to aklog.
My understanding was that this setting was only
needed on the KDCs, which until now, has been
working fine since
On 4 Nov 2011, at 19:14, David R Boldt wrote:
We are seeing 3-5 million entries per day in VolserLog
that looks something like the following:
Thu Nov 3 11:07:22 2011 ReadVnodes: setup 5458/7838205
You should only get that message if you've requested extra logging with the
-log
On 4 Nov 2011, at 21:24, David R Boldt wrote:
What's the most pain-free way to update the options;
can I modify BosConfig and restart volserver?
copy BosConfig to BosConfig.new, make the change in BosConfig.new, then restart
the bosserver. If you're running non-dafs, restarting the
On 31 Oct 2011, at 10:03, Jan Krcmar honza...@gmail.com wrote:
i'm trying to create rw volume on afs1:vicepa and ro volume on afs1:vicepb
the following problem occurs:
This isn't a supported configuration. You can't have a replica volume on a
different partition of the same server. Cheap
On 23 Sep 2011, at 00:28, Derrick Brashear sha...@gmail.com wrote:
On Thu, Sep 22, 2011 at 3:30 PM, Steve Simmons s...@umich.edu wrote:
I've been working on a patch to AFS so that one could add scales to numbers.
The primary goal is to be able to do things like
$ fs setq . 1g
commit
On 22 Sep 2011, at 08:31, Dan Scott danieljamessc...@gmail.com wrote:
No, I haven't, because two developers confirmed that I'd need to migrate:
https://www.redhat.com/archives/freeipa-users/2011-May/msg00250.html
https://www.redhat.com/archives/freeipa-users/2011-May/msg00251.html
If the
On 21 Sep 2011, at 23:08, Dan Scott danieljamessc...@gmail.com wrote:
I have to perform a fairly major upgrade on my Kerberos servers which
authenticate our Openafs cell, which means running with 2 different
kerberos servers, at least for a short while.
Running with two different KDCs, both
On 19 Sep 2011, at 09:45, Staffan Hämälä wrote:
I've been trying to compile the AFS perl module against openAFS 1.4.14 on a
64-bit CentOS 6 Linux machine, but it looks like it's not easy to get it to
work.
Yes, perl-AFS doesn't really work on 64bit platforms. This is because its
library
On 18 Sep 2011, at 11:13, Frank Burkhardt wrote:
Hi Everyone,
I'm currently doing a testrun with Ubuntu Natty + openafs 1.6.0 (Russ
Allbery's Debian package version 1.6.0-1).
When I do
root@myhost fs setserver -vl someserver 1000
I get this message:
This cache manager does not
We're currently maintaining 3 versions of the AFS fileserver - the LWP one, a
normal pthreaded fileserver, and the demand attach fileserver. The normal
pthreaded fileserver has been the default for all of our supported platforms
since the 1.4 release.
I'd like to simplify the build tree, and
On 5 Sep 2011, at 17:12, Berthold Cogel wrote:
Today I tried to install the 1.6.0 packages for fedora from openafs.org.
I've created a repo file and yum install worked as desired for almost
all packages. Only the kernel module was rejected:
This is slightly strange - did you yum install all
To paraphrase a famous AFS saying: The fastest disk access is the one you don't
do.
If your disks are having performance issues (I'd also be interested in seeing
utilisation figures, and the difference between the await and svctm times),
then the easiest way to fix that is to have less
On 29 Aug 2011, at 17:33, Harald Barth wrote:
PS: Some years ago I did some benchmarking of AFS cache performance. I
was astonished how much slower ext2 on ramdisk was compared to
memcache. So the question is, how fast could a AFS disk cache be if it
did not needed to bother with a file
On 24 Aug 2011, at 14:17, Jaap Winius wrote:
But, wouldn't that cause trouble in the mornings when those workstations are
turned on again and a number of them try to re-synchronize their 1GB caches
across my relatively puny 6Mbps WAN links?
Caches aren't revalidated, or resyncronised, upon
It's a wiki. You can edit it.
S.
Sent from my iPhone
On 24 Aug 2011, at 15:20, Jeff Blaine jbla...@kickflop.net wrote:
For read/write data, if the cache is too small, the cache manager is
required to flush data to the file server sooner than it would prefer.
Since many files used today are
On 22 Aug 2011, at 15:07, Rick Cochran r...@cornell.edu wrote:
Although potentially disruptive, these seem logical to me.
Do you know of any others?
The whole of MIT Kerberos has been ripped out and replaced with Heimdal. The
Heimdal API is private. Bits of the old MIT API are supported via
On 11 Aug 2011, at 17:44, Dale Pontius wrote:
I was under the impression that 1.4.14.1 was it, and that 1.6.0 would be the
next release out, which begs the next question.
Will 1.4.15 support the linux-3.0.x kernel series? It was actually
linux-2.6.39 that drove me off of 1.4.14.1.
So, I
On 2 Aug 2011, at 15:09, Derrick Brashear sha...@gmail.com wrote:
#define KRB5_PROG_ETYPE_NOSUPP (-1765328234L)
set the
allow_weak_crypto = yes
option in /Library/Preferences/edu.mit.Kerberos
Firstly, I thought that the 1.6 aklog converted both Kerberos and AFS errors
On 31 Jul 2011, at 21:15, Aaron Knister aar...@umbc.edu wrote:
Make that a vos syncserv not a syncvldb
If I'm understanding your problem correctly you have a single entry for a RO
volume in your VLDB, with the corresponding server long since departed. In this
case, vos delentry should let
I saw a mention of IPv6 support sometime in 2011 in my old email..
It's a work in progress.
There are actually multiple components to achieving IPv6 support in AFS-3.
Different amounts of progress have been made on each of these components...
1/ Refactor the OpenAFS RX API so that IPv6
On 21 Jun 2011, at 17:16, Russ Allbery r...@stanford.edu wrote:
Nicolas Bourbaki ncl.bourb...@gmail.com writes:
1.6.0pre6 made it to the ubuntu ppa repository, however the issue is
still present. It is caused by KDE's akonadi-mysql server trying to bind
to a UNIX socket inside the user's
On 19 Jun 2011, at 21:32, Matt W. Benjamin wrote:
One way you can take deterministic advantage of more memory in today's
fileservers (caveat, if you have the workload) is to increase the number of
callbacks supported. If you're not already setting a million or so, for rw
file service,
On 9 Jun 2011, at 16:06, Andrew Deason wrote:
If you wanted to add recording of timing data, the best place is
probably in afs_read, afs_write, afs_open, etc functions in the
src/afs/VNOPS/, just recording the time spent in each of those
functions.
The Linux cache manager's VFS interface (in
On 7 Jun 2011, at 15:08, Mike Legg mike.l...@u-blox.com wrote:
Many thanks Jeffrey, that was indeed the problem
In the forthcoming 1.6 release, we've fixed this useless error message to say
something a bit more informative, too.
S.___
OpenAFS-info
On 31 May 2011, at 20:01, Dvorkin, Asya dvork...@umdnj.edu wrote:
Currently, we are running our cell on CentOS, but looking into possibly
moving towards Scientific Linux.
Would love to hear your opinions and recommendations.
Being binary rebuilds of the same upstream source distribution,
On 17 May 2011, at 10:54, Mike Legg mike.l...@u-blox.com wrote:
We are looking to evaluate OpenAFS and were wondering where the best place to
get initial version / platform information. We would like to test OpenAFS on
CentOS or Debian but are not sure which versions of the application to
On 18 Apr 2011, at 12:33, Jan Johansson wrote:
Some time ago (in thread
https://lists.openafs.org/pipermail/openafs-info/2011-February/035407.html)
I asked about the client -daemons flag.
Reviewing your original post, it has occurred to me that your problem could be
a symptom of an issue a
Could you run gdb against your kernel module (either openafs.ko or libafs.ko),
and run
list *afs_GetDownD.clone.5+0x1d0
This will let us know exactly where in your kernel module the fault is
occurring.
Cheers,
Simon
___
OpenAFS-info mailing list
On 16 Apr 2011, at 12:42, Ciprian Dorin Craciun wrote:
Have I missed some configuration options, or should I change the
kernel config?
You need CONFIG_DEBUG_KERNEL and CONFIG_DEBUG_INFO in your kernel config to
build modules with debugging symbols.
Cheers,
Simon
Hi,
One of the issues that comes up from time to time is what actually constitutes
a bug worthy of a security advisory. Sometimes this is really clear cut, but in
other areas, in particular in relation to our Unix kernel modules, the dividing
line is significantly less clear. Getting this
On 15 Apr 2011, at 20:43, David Boyes wrote:
A variation of this comment: much of the complexity of deploying a fix is
related to packaging. Investment in simplifying and automating the process of
creating and deploying a new package would probably help somewhat with the
pain level of
On 14 Apr 2011, at 18:20, Jack Neely wrote:
At this point the kernel module compiles but complains about GPL only
symbols:
FATAL: modpost: GPL-incompatible module libafs.ko uses GPL-only
symbol '__init_work'
I don't see anything in git that addresses this. Any ideas?
You kernel
On 12 Apr 2011, at 16:41, Danko Antolovic wrote:
I have the encryption des-cbc-md5 in three other AFS-related keytabs, which
asetkey has been able to process.
Is there a way to narrow down the meaning/origin of that error?
Well, the error is KRB5_KT_NOTFOUND, which means pretty much
(removed openafs-devel)
On 11 Apr 2011, at 10:58, Meisam Mohammadkhani meisam.mohammadkh...@gmail.com
wrote:
The callback mechanism does not guarantee that you immediately see the
changes someone else makes to a file you are using. Your Cache Manager does
not notice the broken callback
On 11 Apr 2011, at 12:12, Meisam Mohammadkhani meisam.mohammadkh...@gmail.com
wrote:
If a client crashes before it closes the file that he changed it before, what
happens to changed data? All of it will lost?
If the application crashes, then the file will still be written to the file
On 11 Apr 2011, at 17:13, Matt W. Benjamin wrote:
The variant of OpenAFS called rxosd has write mirroring, too, I believe
One of the discoveries of the Edinburgh Hackathon was that whilst rxosd has
mirroring, it doesn't have any mechanism for keeping the mirrors in sync. This
means that if a
On 11 Apr 2011, at 16:20, Andrew Deason wrote:
I think I remember looking into this on Linux a long time ago, but I
think I discovered that inotify interaction required GPL-only symbols.
Any idea if that's correct / still the case?
It was correct, but looking at the kernel tree it seems like
On 11 Apr 2011, at 21:11, Derrick Brashear wrote:
And also, to be clear: I was just wondering about the technical details.
I'm pretty sure I wouldn't even use this functionality if we had it, and
I haven't heard a lot of clamoring for it (the search indexing project I
consider to be
On 11 Apr 2011, at 20:47, Russ Allbery wrote:
Simon Wilkinson s...@inf.ed.ac.uk writes:
There is one exception to this behaviour on Unix. If you have opened a
file for writing, and dirtied that file, then the version of the data on
your client remains that at the point the file was dirtied
On 4 Apr 2011, at 22:18, Garrett Wollman wrote:
Over the past few days I have performed several benchmarks comparing
the performance of various OpenAFS server and client configurations.
Thanks for this - it makes for really interesting reading.
The statistic I'm really interested in at
On 5 Apr 2011, at 14:35, John Hascall wrote:
I downloaded openafs-1.6.0pre4 source,
[...]
Then the instructions kinda petered out...
What is your desired end product, and what instructions are you following?
S.
___
OpenAFS-info mailing list
On 5 Apr 2011, at 14:59, John Hascall wrote:
On 5 Apr 2011, at 14:35, John Hascall wrote:
I downloaded openafs-1.6.0pre4 source,
[...]
Then the instructions kinda petered out...
What is your desired end product, and what instructions are you following?
A functioning afs client. The
On 5 Apr 2011, at 15:37, Todd Lewis wrote:
Those two sentences would be a great addition to the src tarball README.
Feel free to add them!
There should be an SRPM available for 1.6.0pre4 from the OpenAFS website,
Should be, but alas, http://dl.openafs.org/dl/openafs/candidate/1.6.0pre4
On 18 Mar 2011, at 10:27, TIARA System Man wrote:
sorry to bother you. now, it is fixed by restoring the previous
kerberos backup. however, i still don't know what is the correct
procedure to add another afs file server with afs key?
You have to have the same afs key on every file and
On 6 Mar 2011, at 03:40, Russ Allbery wrote:
I'm not sure what we were doing before.
The positional I/O code only landed with 335ccb, which is on master and 1.6,
but not on 1.4, which is why we haven't seen this problem before. I suspect
that the correct solution is both to take Ryan's
On 6 Mar 2011, at 03:23, Ryan C. Underwood wrote:
Yes, this was exactly it. Attached patch fixes the issue.
Thanks. Pushed to gerrit as http://gerrit.openafs.org/4152
S.
___
OpenAFS-info mailing list
OpenAFS-info@openafs.org
On 6 Mar 2011, at 10:46, Simon Wilkinson wrote:
On 6 Mar 2011, at 03:23, Ryan C. Underwood wrote:
Yes, this was exactly it. Attached patch fixes the issue.
Thanks. Pushed to gerrit as http://gerrit.openafs.org/4152
This fails on both Irix and AIX, as they don't have preadv64
On 6 Mar 2011, at 13:46, Chaz Chandler wrote:
This fails on both Irix and AIX, as they don't have preadv64 and pwritev64.
I don't have either of those platforms in front of me, so it's not clear
whether the correct thing to do is to just disable positional I/O if
preadv64 and pwritev64
On 6 Mar 2011, at 20:45, Russ Allbery r...@stanford.edu wrote:
Ryan's patch is basically a partial reimplementation of
_FILE_OFFSET_BITS=64 only for that one function.
Actually, we're already carrying this logic to provide 64bit variants of all of
the other file manipulation functions we use.
On 5 Mar 2011, at 21:31, Ryan C. Underwood wrote:
I also looked for any read, seek, or stat call that returned negative.
No luck. It seems like all the threads are being captured...
Strange.
Another thing you could try is (if this is a test system) attach to the
fileserver process with
On 4 Mar 2011, at 06:32, Ryan C. Underwood wrote:
I am having trouble copying a large file (6GB) from a volume located on
one server to a volume located on another server. After about 2GB
(2147295232 bytes to be exact), the volume gets offlined and marked
needs salvage. I have reproduced
1 - 100 of 410 matches
Mail list logo
