Weird pveperf results

[jm.trillo]

I'm a newbie in proxmox, I installed it in a machine to do some tests and see what we can do with it. Pveperf results are giving me results I dont know they're if they're normal.

If I just run pveperf, I get results that seem to be ok, considering it's using only a raid 1:

pveperf
CPU BOGOMIPS: 57596.88
REGEX/SECOND: 1498626
HD SIZE: 19.38 GB (/dev/sda1)
BUFFERED READS: 169.84 MB/sec
AVERAGE SEEK TIME: 7.58 ms
FSYNCS/SECOND: 1150.48
DNS EXT: 34.08 ms
DNS INT: 2.77 ms (ovh.net)

However, if I run pveperf over /var/lib/vz (the lvm volume) I get much lower fsync results:

pveperf /var/lib/vz
CPU BOGOMIPS: 57596.88
REGEX/SECOND: 1547928
HD SIZE: 968.87 GB (/dev/mapper/pve-data)
BUFFERED READS: 187.72 MB/sec
AVERAGE SEEK TIME: 12.75 ms
FSYNCS/SECOND: 167.90
DNS EXT: 31.41 ms
DNS INT: 3.00 ms (ovh.net)


The machine has a single raid 1 (hardware, with write cache on), with a primary partition in /, the LVM in /var/lib/vz and swap. I've tried reinstalling it all without LVM and this dont seem to happen, i got consinstently 800-900 fsyncs. So my questions are,

1) is this the normal behaviour i should expect with pveperf and lvm or am I losing that much fsync performance? After googling a bit I have not found a similar case.

2) If this is a weird hardware/kernel compatibility problem, would I be better off reinstalling everything without lvm and forgettting about snapshots features and so on? The machine is a ovh EG 64G btw,

Thanks in advance for any help.

PD: Sorry if this gets duplicated, having some problems posting. EDIT: Ah the " you need approval" message flashed too fast the first time I tried to post. Sorry

 

[tom]

no, pveperf results should not differ with lvm. but make sure that your server is NOT under load when you run pveperf.

what filesystem do you use? post:

cat /proc/mounts

 

[michu]

Hi,

Check filesystem type on / and /var/lib/vz and check mount options: cat /etc/fstab

Regards,
michu

 

[jm.trillo]

Here:

# <file system> <mount point> <type> <options> <dump> <pass>
/dev/sda1 / ext3 errors=remount-ro 0 1
/dev/sda2 swap swap defaults 0 0
/dev/pve/data /var/lib/vz ext3 defaults 1 2
proc /proc proc defaults 0 0
sysfs /sys sysfs defaults 0 0


System isnt under any load and i've repeated the tests serveral times. I'm using ext3

EDTI: and cat /proc/mounts

cat /proc/mounts
none /sys sysfs rw,nosuid,nodev,noexec,relatime 0 0
none /proc proc rw,nosuid,nodev,noexec,relatime 0 0
none /dev devtmpfs rw,relatime,size=32961120k,nr_inodes=8240280,mode=755 0 0
none /dev/pts devpts rw,nosuid,noexec,relatime,gid=5,mode=620,ptmxmode=000 0 0
/dev/sda1 / ext3 rw,relatime,errors=remount-ro,barrier=0,data=ordered 0 0
tmpfs /lib/init/rw tmpfs rw,nosuid,relatime,mode=755 0 0
tmpfs /dev/shm tmpfs rw,nosuid,nodev,relatime 0 0
fusectl /sys/fs/fuse/connections fusectl rw,relatime 0 0
none /sys/kernel/config configfs rw,relatime 0 0
/dev/fuse /etc/pve fuse rw,nosuid,nodev,relatime,user_id=0,group_id=0,default_permissions,allow_other 0 0
/dev/mapper/pve-data /var/lib/vz ext3 rw,relatime,errors=continue,barrier=0,data=ordered 0 0
beancounter /proc/vz/beancounter cgroup rw,relatime,blkio,name=beancounter 0 0
container /proc/vz/container cgroup rw,relatime,freezer,devices,name=container 0 0
fairsched /proc/vz/fairsched cgroup rw,relatime,cpuacct,cpu,cpuset,name=fairsched 0 0

 

[jm.trillo]

I'm reinstalling manualy over debian and creating the new lvm manually etc to check if it's some error in the ovh installer. If this keep happening, will I get similar performance using only ext3? I can live without snapshots for the moment as long as the performance is good.

 

[tom]

I suggest you test your harddrive (also check smart status) - is this a single drive?

 

[jm.trillo]

2x 2TB SATA3 on a hardware RAID 1, i'll run some more hardware tests once i've finished this install. Since the problem only shows up on LVM I didnt check hardware much.

 

[jm.trillo]

It looks like it's hardware after all. I've installed again, keeping the same default layout ( one partition /, another /var/lib/vz), except the one in /var/lib/vz is ext3 instead of lvm. The fsync count still goes down on that partition, so there must be faulty sectors or something on the hdd.