similar to: LVM Snapshots: COW Internals

On Wed, Dec 02, 2015 at 08:53:39PM +0100, Axel Glienke wrote: > Creating snapshot: > > [root at lvmtest ~]# lvcreate -L5G -s -n root_snap /dev/centos/root > Reducing COW size 5,00 GiB down to maximum usable size 2,94 GiB. > Logical volume "root_snap" created. > [root at lvmtest ~]# lvs > LV VG Attr LSize Pool Origin Data% Meta% Move > Log

Hi there, I got really tired of false hangups, specially when someone calls from SIP/IAX/whatever to PSTN, which makes no sense in using busydetect in the zap channel, as the caller will eventually hear the busy tones and hangup, causing the zap channel to be "freed" as well... After suggesting that the busy detect be enabled only for inbound calls, and finding other people that think

in journalctl i found: modprobe: FATAL: Module dm-snapshot not found ... Can't process LV root_snap: snapshot target support missing from kernel Zitat von Tru Huynh <tru at centos.org>: > On Wed, Dec 02, 2015 at 08:53:39PM +0100, Axel Glienke wrote: >> Creating snapshot: >> >> [root at lvmtest ~]# lvcreate -L5G -s -n root_snap /dev/centos/root >> Reducing

Creating snapshot: [root at lvmtest ~]# lvcreate -L5G -s -n root_snap /dev/centos/root Reducing COW size 5,00 GiB down to maximum usable size 2,94 GiB. Logical volume "root_snap" created. [root at lvmtest ~]# lvs LV VG Attr LSize Pool Origin Data% Meta% Move Log Cpy%Sync Convert root centos owi-aos--- 2,93g root_snap centos swi-a-s--- 2,94g

Hi all, So I have an XFS file system within LVM which has an external log. My mount option in FSTAB is; /dev/vg_spock_data/lv_data /data xfs logdev=/dev/sdc1,nobarrier,logbufs=8,noatime,nodiratime 1 1 All is well no issues and very fast. Now I'd like to snapshot this bad boy and then run rsnapshot to create a few days backup. A snapshot volume is created w/o issue; lvcreate -L250G -s

sorry: centos7 "fresh" minimal with a actual update cat proc/cmdline: BOOT_IMAGE=/vmlinuz-3.10.0-229.20.1.el7.x86_64 root=/dev/mapper/centos_root ro rd.lvm.lv=centos/root rd.lvm.lv=centos/swap crashkernel=auto rhgb quiet LANG=de_DE.UTF-8 systemd.debug ls /dev/mapper: centos-swap control Zitat von Gordon Messmer <gordon.messmer at gmail.com>: > On 12/02/2015 11:09 AM,

Hi, using btrfs with LVM snapshots seems to be confusing /proc/mounts After mounting a snapshot of an original filesystem, the devicename of the original filesystem is overwritten with that of the snapshot in /proc/mounts. Steps to reproduce: arnd@kallisto:/mnt$ sudo mount /dev/vg0/original /mnt/original [ 107.041432] device fsid 5c3e8ca2-da56-4ade-9fef-103a6a8a70c2 devid 1 transid 4

Two days ago, I was talking on the phone from the FXO, to a SIP phone. After some time (like 1h30m), all of a sudden, there's a huge noise, like a buzz... Really loud. So I hungup, and called my asterisk box again... All I could hear was that sound. Someone called me from the internet, and as Asterisk dialed the FXO, all she heard was that noise too. So, I logged in my Asterisk server,

Hi everyone! I''ve been reading this list for a while and I''m thinking that the best way to do backups is using LVM snapshots as discussed here several times. The problem is that when i use a LVM logical volume as windows disk, windows writes the partition table onto the LV so the LV itself is not a partition - it''s a whole disk. lvm/device-mapper does not create the

Hi there, I''ve read the paper on Xen live migration, and it shows some very impressive figures, like 165ms downtime on a running web server, and 50ms for a quake3 server. I installed CentOS 5 on 2 servers, each with 2x Xeon E5335 (quad-core), 2x Intel 80003ES2LAN Gb NICs. Then I installed 2 DomUs, also with CentOS 5. One NIC is connected to the LAN (on the same switch and VLAN), the

Hi, I just installed a CentOS 6.5 System with the intention of using thinly provisioned snapshots. I created the volume group, a thinpool and then a logical volume. All of that works fine but when I create a snapshot "mysnap" then the snapshot volume gets displayed in the "lvs" output with the correct information but apparently no device nodes are created under

I had this idea, and after looking for something like this already in progress, I found another guy who tried to start it... But I was unable to contact him, and his project seems to be dead. But, I believe it is possible, and I wanted to know the opinion of the experienced... So, let's go: I got an asterisk server setup to receive free calls from US to Brazil. The problem is that at my work,

Hi there, I was looking for a GSM gateway, but I didn't find any prices... Anybody knows how much one of these costs? And, is it possible to use an amateur radio with asterisk? Thanks, Marconi.

Just discovered that on CentOS 4.1 when you issue the command: lvcreate -s -n temp -L2G /dev/vgname/lvname to create a snapshot, you get the error message: snapshot: Required device-mapper target(s) not detected in your kernel I've checked and recheck my spelling and tried using the device created under /dev/mapper. I've tried it on both x86_64 and on i386. Works as expected under

I've been experimenting with LVM and snapshots under Centos 4.4 recently, and have found performance to be horrible while a snapshot of a logical volume exists. I expected some performance degradation with a snapshot in place due to my understanding of the copy-on-write behavior, but my system becomes almost unusable when writing to my system with a snapshot in place. For example, I can

We utilize LVM snapshots for some periodic maintenance. They're manually created and, usually, manually destroyed. But not always. So there's now a nightly script monitoring for open snapshots. Which raises the question of when a given snapshot was created. Absent good practices of, say, using sudo to create snapshots (leaving a /var/log/secure message), is there any reasonably

Hello, I have a Centos 6.2 clsuter with a CLVM partition on which I have a GFS2 file system. The problem rises when I make a snapshot from my FC NetAPP FAS2020. After I make the snapshot (it is a rw snapshot) of my LUN, I am not able to mount it from any of my cluster nodes, because the Physical Volume is seen two times one time on the standard LVM partition and the other time on the snapshot

Not a centos specific question here, but if anyone can save me from shooting myself in the foot would appreciate any pointers.... I have an older centos 4.7 box that I recently replaced with a newer one running centos 5.3. I'd like to add the hard disk from the older one to the newer one before I scrap it for good. I don't care about the data on it, would just like the extra drive

The cow filter doesn't support trimming and we should assume the underlying plugin is fully allocated too. Note that both of these limitations might be lifted with a more advanced filter implementation. However we ought to support this in future because xz files contain sparseness information, so add a note. --- filters/cow/cow.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-)

I want to eliminate the COW feature on all of my OS files. It is a nice feature for user files, but I don''t see a clear benefit for the actual OS files. And I suspect that COW induced fragmentation is causing or aggravating problems with my system including the boot open_ctree problem. I had planned to recursively chattr these files to "nodatacow" status but then I ran