krisdigitx
asked on
using bacula NFS mount to a NAS server
hi,
I just wanted to get some opinions of configuring this setup, I have a 7TB server which I mount as a NFS share to a bacula server(ubuntu 8.10) for backups. The NFS/NAS server is connected to bacula backtoback by a ethernet connection.
The NAS server(debian) is a software RAID 5 with 6 x 1.5TB IDE/SATA disks.
Is there any constraints in the long run? will there be performance issues and which areas need to be upgraded?
any suggestions will be helpful
cheers!
I just wanted to get some opinions of configuring this setup, I have a 7TB server which I mount as a NFS share to a bacula server(ubuntu 8.10) for backups. The NFS/NAS server is connected to bacula backtoback by a ethernet connection.
The NAS server(debian) is a software RAID 5 with 6 x 1.5TB IDE/SATA disks.
Is there any constraints in the long run? will there be performance issues and which areas need to be upgraded?
any suggestions will be helpful
cheers!
nabeelmoidu
I think the main points would be the cache on the device, your controller speed and network connectivity.if you can give more details about your hardware we can help you out. Its not very likely you might get people who actually worked on bacula here.
Copy 7G over gigabit wire will take 15 hours
7G off the 50MB/s disk (common) will take 30 hours
Striping (raid0)disks in backup server is absolute must. SW breed is OK
Benchmark like iozone will tell you better.
gzip -1 compression can help
teaming network adapters is absolute must for NFS server to keep serving people while backing up (or double backup time)
backup is not dependent on disk cache - it reads all the physical media, it is very likely to hit disk errors, so use "smartmontools" religiously.
You might need to rebuild gzip to your CPU so that it goes up to speed.
PS it is highly likely to see bacula in academic settings.
7G off the 50MB/s disk (common) will take 30 hours
Striping (raid0)disks in backup server is absolute must. SW breed is OK
Benchmark like iozone will tell you better.
gzip -1 compression can help
teaming network adapters is absolute must for NFS server to keep serving people while backing up (or double backup time)
backup is not dependent on disk cache - it reads all the physical media, it is very likely to hit disk errors, so use "smartmontools" religiously.
You might need to rebuild gzip to your CPU so that it goes up to speed.
PS it is highly likely to see bacula in academic settings.
ASKER
this is the server configuration,
smartmontools is already installed and running,
smartmontools is already installed and running,
lf01:~# cat /proc/cpuinfo
processor : 0
vendor_id : AuthenticAMD
cpu family : 15
model : 65
model name : Dual-Core AMD Opteron(tm) Processor 2214
stepping : 3
cpu MHz : 2211.442
cache size : 1024 KB
physical id : 0
siblings : 2
core id : 0
cpu cores : 2
apicid : 0
initial apicid : 0
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt rdtscp lm 3dnowext 3dnow pni cx16 lahf_lm cmp_legacy svm extapic cr8_legacy
bogomips : 4426.26
clflush size : 64
power management: ts fid vid ttp tm stc
processor : 1
vendor_id : AuthenticAMD
cpu family : 15
model : 65
model name : Dual-Core AMD Opteron(tm) Processor 2214
stepping : 3
cpu MHz : 2211.442
cache size : 1024 KB
physical id : 0
siblings : 2
core id : 1
cpu cores : 2
apicid : 1
initial apicid : 1
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt rdtscp lm 3dnowext 3dnow pni cx16 lahf_lm cmp_legacy svm extapic cr8_legacy
bogomips : 4422.70
clflush size : 64
power management: ts fid vid ttp tm stc
processor : 2
vendor_id : AuthenticAMD
cpu family : 15
model : 65
model name : Dual-Core AMD Opteron(tm) Processor 2214
stepping : 3
cpu MHz : 2211.442
cache size : 1024 KB
physical id : 1
siblings : 2
core id : 0
cpu cores : 2
apicid : 2
initial apicid : 2
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt rdtscp lm 3dnowext 3dnow pni cx16 lahf_lm cmp_legacy svm extapic cr8_legacy
bogomips : 4422.71
clflush size : 64
power management: ts fid vid ttp tm stc
processor : 3
vendor_id : AuthenticAMD
cpu family : 15
model : 65
model name : Dual-Core AMD Opteron(tm) Processor 2214
stepping : 3
cpu MHz : 2211.442
cache size : 1024 KB
physical id : 1
siblings : 2
core id : 1
cpu cores : 2
apicid : 3
initial apicid : 3
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt rdtscp lm 3dnowext 3dnow pni cx16 lahf_lm cmp_legacy svm extapic cr8_legacy
bogomips : 4422.72
clflush size : 64
power management: ts fid vid ttp tm stc
lf01:~#
lf01:~# free
total used free shared buffers cached
Mem: 3632608 301664 3330944 0 1180 191116
-/+ buffers/cache: 109368 3523240
Swap: 3903752 0 3903752
lf01:~#
dmesg is of more help.
if you run amd64 system it is possible that gzip is optimal already.
could you please run iozone to see if disks are faster than network and if you can gain from compression?
if you run amd64 system it is possible that gzip is optimal already.
could you please run iozone to see if disks are faster than network and if you can gain from compression?
ASKER
i ran this test over the NFS link
/var/www/vhosts is mounted from the NFS server(AMD 7TB server)
/var/www/vhosts is mounted from the NFS server(AMD 7TB server)
Iozone: Performance Test of File I/O
Version $Revision: 3.327 $
Compiled for 32 bit mode.
Build: linux
Contributors:William Norcott, Don Capps, Isom Crawford, Kirby Collins
Al Slater, Scott Rhine, Mike Wisner, Ken Goss
Steve Landherr, Brad Smith, Mark Kelly, Dr. Alain CYR,
Randy Dunlap, Mark Montague, Dan Million, Gavin Brebner,
Jean-Marc Zucconi, Jeff Blomberg, Benny Halevy,
Erik Habbinga, Kris Strecker, Walter Wong, Joshua Root.
Run began: Mon Jan 11 11:11:26 2010
Excel chart generation enabled
Record Size 4 KB
File size set to 102400 KB
Command line used: /opt/iozone/bin/iozone -R -l 10 -u 10 -r 4k -s 100m -F /var/www/vhosts/f1 /var/www/vhosts/f2 /var/www/vhosts/f3 /var/www/vhosts/f4 /var/www/vhosts/f5 /var/www/vhosts/f6 /var/www/vhosts/f7 /var/www/vhosts/f8 /var/www/vhosts/f9 /var/www/vhosts/10 output.txt
Output is in Kbytes/sec
Time Resolution = 0.000001 seconds.
Processor cache size set to 1024 Kbytes.
Processor cache line size set to 32 bytes.
File stride size set to 17 * record size.
Min process = 10
Max process = 10
Throughput test with 10 processes
Each process writes a 102400 Kbyte file in 4 Kbyte records
Children see throughput for 10 initial writers = 915439.45 KB/sec
Parent sees throughput for 10 initial writers = 24475.38 KB/sec
Min throughput per process = 573.60 KB/sec
Max throughput per process = 414805.00 KB/sec
Avg throughput per process = 91543.95 KB/sec
Min xfer = 420.00 KB
Children see throughput for 10 rewriters = 1085968.31 KB/sec
Parent sees throughput for 10 rewriters = 73246.33 KB/sec
Min throughput per process = 0.00 KB/sec
Max throughput per process = 690697.44 KB/sec
Avg throughput per process = 108596.83 KB/sec
Min xfer = 0.00 KB
Children see throughput for 10 readers = 110983.26 KB/sec
Parent sees throughput for 10 readers = 105681.77 KB/sec
Min throughput per process = 9486.93 KB/sec
Max throughput per process = 15482.62 KB/sec
Avg throughput per process = 11098.33 KB/sec
Min xfer = 64336.00 KB
Children see throughput for 10 re-readers = 2314355.01 KB/sec
Parent sees throughput for 10 re-readers = 1982440.32 KB/sec
Min throughput per process = 8974.01 KB/sec
Max throughput per process = 1037539.81 KB/sec
Avg throughput per process = 231435.50 KB/sec
Min xfer = 924.00 KB
Children see throughput for 10 reverse readers = 1539350.92 KB/sec
Parent sees throughput for 10 reverse readers = 1482131.59 KB/sec
Min throughput per process = 34.76 KB/sec
Max throughput per process = 786095.81 KB/sec
Avg throughput per process = 153935.09 KB/sec
Min xfer = 4.00 KB
Children see throughput for 10 stride readers = 1288892.79 KB/sec
Parent sees throughput for 10 stride readers = 1216024.05 KB/sec
Min throughput per process = 28681.06 KB/sec
Max throughput per process = 659254.19 KB/sec
Avg throughput per process = 128889.28 KB/sec
Min xfer = 4544.00 KB
Children see throughput for 10 random readers = 1522220.33 KB/sec
Parent sees throughput for 10 random readers = 1467732.92 KB/sec
Min throughput per process = 724.64 KB/sec
Max throughput per process = 825160.56 KB/sec
Avg throughput per process = 152222.03 KB/sec
Min xfer = 92.00 KB
Children see throughput for 10 mixed workload = 1137702.12 KB/sec
Parent sees throughput for 10 mixed workload = 26290.57 KB/sec
Min throughput per process = 1133.19 KB/sec
Max throughput per process = 684931.06 KB/sec
Avg throughput per process = 113770.21 KB/sec
Min xfer = 172.00 KB
Children see throughput for 10 random writers = 564308.42 KB/sec
Parent sees throughput for 10 random writers = 1153.03 KB/sec
Min throughput per process = 464.65 KB/sec
Max throughput per process = 456997.94 KB/sec
Avg throughput per process = 56430.84 KB/sec
Min xfer = 436.00 KB
Children see throughput for 10 pwrite writers = 559548.02 KB/sec
Parent sees throughput for 10 pwrite writers = 26084.40 KB/sec
Min throughput per process = 1613.10 KB/sec
Max throughput per process = 339352.72 KB/sec
Avg throughput per process = 55954.80 KB/sec
Min xfer = 40.00 KB
Children see throughput for 10 pread readers = 110567.23 KB/sec
Parent sees throughput for 10 pread readers = 104486.70 KB/sec
Min throughput per process = 9584.21 KB/sec
Max throughput per process = 16444.08 KB/sec
Avg throughput per process = 11056.72 KB/sec
Min xfer = 59728.00 KB
"Throughput report Y-axis is type of test X-axis is number of processes"
"Record size = 4 Kbytes "
"Output is in Kbytes/sec"
" Initial write " 915439.45
" Rewrite " 1085968.31
" Read " 110983.26
" Re-read " 2314355.01
" Reverse Read " 1539350.92
" Stride read " 1288892.79
" Random read " 1522220.33
" Mixed workload " 1137702.12
" Random write " 564308.42
" Pwrite " 559548.02
" Pread " 110567.23
iozone test complete.
-s should exceed all the imaginable caches 2-10fold.
-s8G or more looks like a hardware test.
On the other hand - results are quite good for memory cache - they exceed gigabit network, and this means no need for nfs tuning...
-s8G or more looks like a hardware test.
On the other hand - results are quite good for memory cache - they exceed gigabit network, and this means no need for nfs tuning...
ASKER
so you mean, bacula would work seamlessly with this config?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
I have configured bacula to use the NFS mount from the storage server, works pretty cool now.. i will keep a check...