2012/9/27 Rodrigo Abrantes Antunes
<rodrigoantunes AT pelotas.ifsul.edu DOT br>
It is about 1MB/s.
Where should I start looking?
In various places. :)
During a backup I can see bacula-sd using 100% cpu, memory usage is normal. My system is a vmware virtual machine with a cpu with two 2,5ghz cores and 2gb ram. The vm have a separate partition for the volumes using ext4 and mounted via RDM in a fc storage array with raid6. I tried mapping more cores and more ram for the vm with the same result so I don't think the problem is lack of resources.
Thanks.
Director: 5.0.1-1ubuntu1
Storage: 5.0.1-1ubuntu1
It is fairly old version of Bacula. Is it possible to upgrade?
FD: 5.0.1-1ubuntu1 (some clients have lower version)
Database: mysqI
OS: Ubuntu 10.04.4 x64 Server
FC Storage 4 GBits/s.
All my network is Gigabit Ethernet.
First you have to check if your storage server is able to write data to disk fast enough. You can do a test with a simple dd if=/dev/zero of=/your-storage-filesystem bs=65536.
Second you have to check if you set Spool Attributes = yes in your job, if not change it. You'll have to setup a spool directory for your device too.
Next check a network speed. What is your RTT between client and storage daemon? I know it is a 1Gbit, but it is possible that you cannot utilize so much. You can test with a simple ftp client. Is it possible that you have a Firewall between both servers?
Last you have to check a client utilization and filesystem read speed. How many small files you have to backup? Every new file to backup has its own overhead. Bacula has to find this file, stat it, open it, read data from it (with 64k block) and finally close it. If you have a 10M files of 1kB then your backup will be much slower then 1 file of 10GB. You have to be aware of it.
One question: can vmware tools have any influence in this case? From what I know vmware tools is mainly for better graphics performance isn't it? I din't installed it.
It is possible that without vmware tools you are using an emulated storage and network I/O instead of much faster virtualized.
best regards