similar to: --o-direct option

On Fri, Aug 07, 2020 at 07:53:13AM -0500, Eric Blake wrote: > >$ free -m; time ./nbdkit file /var/tmp/random fadvise=sequential cache=none --run 'qemu-img convert -n -p -m 16 -W $nbd "json:{\"file.driver\":\"null-co\",\"file.size\":\"1E\"}"' ; free -m ; cachestats /var/tmp/random > > Hmm - the -W actually says that qemu-img is

https://bugzilla.samba.org/show_bug.cgi?id=7876 Summary: please implement o_direct Product: rsync Version: 3.1.0 Platform: Other OS/Version: Linux Status: NEW Severity: normal Priority: P3 Component: core AssignedTo: wayned at samba.org ReportedBy: costinel at gmail.com QAContact:

You can use these flags as described in the manual page to optimize access patterns, and to get better behaviour with the page cache in some scenarios. For my testing I used the cachedel and cachestats utilities written by Julius Plenz (https://github.com/Feh/nocache). I started with a 32 GB file of random data on a machine with about 32 GB of RAM. At the beginning of the test I evicted the

On Fri, Aug 07, 2020 at 05:29:24PM +0300, Nir Soffer wrote: > On Fri, Aug 7, 2020 at 5:07 PM Richard W.M. Jones <rjones@redhat.com> wrote: > > These ones? > > https://www.redhat.com/archives/libguestfs/2020-August/msg00078.html > > No, we had a bug when copying image from glance caused sanlock timeouts > because of the unpredictable page cache flushes. > > We

On Fri, Aug 07, 2020 at 04:43:12PM +0300, Nir Soffer wrote: > On Fri, Aug 7, 2020, 16:16 Richard W.M. Jones <rjones@redhat.com> wrote: > > I'm not sure if or even how we could ever do a robust O_DIRECT > > > > We can let the plugin an filter deal with that. The simplest solution is to > drop it on the user and require aligned requests. I mean this is very error

The new cache=none mode is misleading since it does not avoid usage of the page cache. When using shared storage, we may get stale data from the page cache. When writing, we flush after every write which is inefficient and unneeded. Rename the cache modes to: - writeback - write complete when the system call returned, and the data was copied to the page cache. - writethrough - write completes

On Sat, Aug 08, 2020 at 02:14:22AM +0300, Nir Soffer wrote: > On Fri, Aug 7, 2020 at 5:36 PM Richard W.M. Jones <rjones@redhat.com> wrote: > > > > On Fri, Aug 07, 2020 at 05:29:24PM +0300, Nir Soffer wrote: > > > On Fri, Aug 7, 2020 at 5:07 PM Richard W.M. Jones <rjones@redhat.com> wrote: > > > > These ones? > > > >

This is a quick hack to experiment with parallel threading model in the python plugin. Changes: - Use aligned buffers to make it possible to use O_DIRECT. Using parallel I/O does not buy us much when using buffered I/O. pwrite() copies data to the page cache, and pread() reads data from the page cache. - Disable extents in the file plugin. This way we can compare it with the python

I'm having the hardest time trying to come up with the optimal configuration with NSS Winbind support. I want it to work right offline. That is, name lookups shouldn't take 30 minutes to time out or lock the system up. And if the name lookup is for a local name, I want Winbind to be 100% out of hte picture. I've tried this, without much luck: passwd: compat [SUCCESS=return] winbind

I am using rsync on a Linux box, connecting to a Cygwin SSH server which has Rsync installed, and pulling some files to the Linux box. The transfer locks up. The rsync process on the Windows box hangs. A single hidden file in progress exists. It is an excel file, with 0 bytes. It locks up on this file consistently. I can open it, read it, delete it, using Cygwin tools, on Windows. It is an Excel

While going through an old todo list I found that these patches had fallen by the way-side. About a year ago I initiated a discussion[1] with the Linux kernel folks regarding the lack of any useable fadvise support on the kernel side. As a result, I was observing extremely poor performance on my server after backup as executable pages were being swapped out in favor of data waiting to be flushed

Hi. I'm having trouble with nut-2.0.3 on a linux kernel 2.4.32 machine and a MGE Evolution 3000. The problem is that I'm getting the following message, repeatidly, on dmesg: usbdevfs: usb_submit_urb returned -22 (-22 means -EINVAL) I traced the problem with newhidups -DDDDDD auto and found out that the -22 error appears in a line "none (-22)" of the output as: ----------

Currently, default behavior of O_DIRECT writes was allowing concurrent writing among nodes, no cluster coherency guaranteed (no EX locks was taken), it hurts buffered reads on other nodes by reading stale data from cache. The new mount option introduce a chance to choose two different behaviors for O_DIRECT writes: * coherency=full, as the default value, will disallow concurrent

We're using RMAN to back up our 9.2 RAC database to an OCFS v1 volume. We have an existing shell script that we use for copying files from disk to tape via tar, one file at a time. (Don't ask why. It's a legacy script. Long story.) We're tweaking this script to use --o_direct when tarring the file to tape and that seems to be working fine: # tape device is /dev/nst0 $ tar

Hello everyone, I am learning and evaluating a glusterfs for film/video editing facilities. Some major film/video editing realtime applications are using the O_DIRECT file access for video/audio data files. The GLFS client via fuse mechanism is disallow the open file with O_DIRECT flag. I made a little sample code for read a file with O_DIRECT flag, and tried open the files on GLFS volumes. It

Does samba 3.0.23c not support the use of O_DIRECT? When I try to open an smbmount'd file using O_DIRECT, I get EINVAL. I am able to use O_DIRECT with no problems on a block device and nfs mounts, so I know the kernel supports it. samba: 3.0.23c kernel: 2.6.9-42.0.3.EL (32-bit) I am using the below code for my test. smb fails on open(2). #include <fcntl.h> #include

Hi We currently use rsync for various jobs at our company. We are now looking at using it to create an offsite synchonised copy of an Oracle 10g RAC archive logs area. The source area is on Oracle OCFS filesystem. The OCFS filesystem requires all reads/writes to be performed with the O_DIRECT option, thus bypassing cache. Oracle provide an updated coreutils package which includes the

Wim, Below are two problems I found in testing the newly released ocfs version. Just for your information. Gr, Robert. - Copying of multiple files gives errors [oracle@prac01 test]$ cp --o_direct -R ./a2/* ./backup/a2/. cp: writing `./backup/a2/./ccdata.dbf.bck': Invalid argument cp: writing `./backup/a2/./ccindex.dbf': Invalid argument cp: writing `./backup/a2/./control02.ctl':

hi list, i have a little problem with 2-node RAC using OCFS. the application running on this cluster does heavily index-based accesses. the data volumes are SAN volumes connected by fibrechannel. the throughput does not exceed 10mb/s, average is 7-8 mb/s. i've used 'iostat -x' and got rkB/s=8000 while %util=100% (device was saturated) from kernel's POV. i did some

In config file, ''*type=ioemu*'' is added with vif . Does it make the difference if we don''t add. Because i did not notice any difference in para as well as full virtulization. Same as in case of exporting the disk to Guest dom. Is it just a placeholder or else ? Second thing, I am running RHEL3 as a HVM DomU on xen-3.0.4. Every thing works fine, but i had a problem to