May 30, 201313 yr When I enable Use Cache Disk on the user share setting and then copy the files on the user shares. The write speed is quite slow (I can see it copied to Cache disk). However, when I copy directly on the Cache disk - the write speed is a lot faster. What happen?
June 2, 201313 yr Author Attach a syslog attached. SMART reports on which disk? If you understand me correctly, when I copy the files to Cache disk directly it is a lot faster (I get like 60-70MB/sec). When I copy files on User Shares (with Use cache disk enabled) - it will copy to cache disk but slow transfer speed like 35MB/sec syslog.zip
June 3, 201313 yr What are the system specifications? Does the issue occur without any add-ons? See my sig to disable add-ons for testing.
June 3, 201313 yr Author What are the system specifications? Does the issue occur without any add-ons? See my sig to disable add-ons for testing. I have disabled the add-ons completely and restarted the server. The speed issue is the same. Its HP ProLiant MicroServer N36L 1.3GHz AMD Athlon II Neo N36L 4GB RAM Running ESXi
June 3, 201313 yr Author Please attach a new syslog. Attached a new syslog with all add-ons disabled. syslog.txt
June 4, 201313 yr I have exactly the same issue, slow write of appr. 35MB when files are copied from Windows7 to a user share with cache disk enabled. But when I copy directly to the cache disk, the transfer speed is 100MB. I tied it also with the latest unraid version 5.0 rc13. No additional plugins are installed and my system is a HP40L with 4M Ram. What is the difference between the 2 options "use cache disk=YES" and "use cache disk=ONLY"
June 4, 201313 yr What is the difference between the 2 options "use cache disk=YES" and "use cache disk=ONLY" The YES option means that files are temporarily stored on the Cache disk (space permitting) and later moved to the main disk shares when the 'mover' application runs overnight. The ONLY option means that the files must be put onto the Cache disk and never moved to user shares on the main disks. This is commonly used for things like app files that should always be left in the same place.
June 4, 201313 yr What does the command top show when copying to cache-enabled share and direct to cache? Is the system running in a virtual machine? Try bare metal for testing. Your hardware should not have this issue with a stock bare metal install.
June 4, 201313 yr Author What does the command top show when copying to cache-enabled share and direct to cache? Is the system running in a virtual machine? Try bare metal for testing. Your hardware should not have this issue with a stock bare metal install. This is what top command show when it start coping to a share with cache-enabled (It would not go over 60% smbd CPU): top - 17:34:08 up 19:37, 1 user, load average: 0.47, 0.14, 0.08 Tasks: 73 total, 2 running, 71 sleeping, 0 stopped, 0 zombie Cpu(s): 9.1%us, 29.2%sy, 0.0%ni, 53.2%id, 1.5%wa, 0.0%hi, 7.0%si, 0.0%st Mem: 2074448k total, 2022832k used, 51616k free, 29568k buffers Swap: 0k total, 0k used, 0k free, 1924400k cached PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 1495 admin 20 0 16700 4532 3392 R 53 0.2 1:06.70 smbd 1223 root 20 0 47780 1476 572 S 49 0.1 1:57.65 shfs 465 root 20 0 0 0 0 S 2 0.0 0:06.58 kswapd0 2149 root 20 0 0 0 0 S 1 0.0 0:00.75 flush-8:64 158 root 20 0 0 0 0 S 1 0.0 0:02.34 sync_supers 2147 root 20 0 2340 1036 820 R 0 0.0 0:00.18 top 1 root 20 0 828 280 240 S 0 0.0 0:13.68 init 2 root 20 0 0 0 0 S 0 0.0 0:00.02 kthreadd 3 root 20 0 0 0 0 S 0 0.0 0:00.19 ksoftirqd/0 Copying direct to cache disk: top - 17:37:27 up 19:40, 1 user, load average: 0.39, 0.20, 0.12 Tasks: 73 total, 2 running, 71 sleeping, 0 stopped, 0 zombie Cpu(s): 7.8%us, 32.8%sy, 0.0%ni, 49.7%id, 0.0%wa, 0.0%hi, 9.8%si, 0.0%st Mem: 2074448k total, 1924724k used, 149724k free, 34120k buffers Swap: 0k total, 0k used, 0k free, 1812508k cached PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 1495 admin 20 0 16532 4380 3400 R 88 0.2 1:35.74 smbd 2149 root 20 0 0 0 0 S 3 0.0 0:02.27 flush-8:64 1 root 20 0 828 280 240 S 0 0.0 0:13.69 init 2 root 20 0 0 0 0 S 0 0.0 0:00.02 kthreadd 3 root 20 0 0 0 0 S 0 0.0 0:00.19 ksoftirqd/0 4 root 20 0 0 0 0 S 0 0.0 0:01.04 kworker/0:0 6 root RT 0 0 0 0 S 0 0.0 0:00.01 migration/0 Yes running unRAID on ESXi. Array Disk and Cache disk are RDM. Don't think it is possible to try bare metal for testing when the disks already setup as RDM?
June 4, 201313 yr PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 1495 admin 20 0 16700 4532 3392 R 53 0.2 1:06.70 smbd 1223 root 20 0 47780 1476 572 S 49 0.1 1:57.65 shfs shfs is the user share system. There does not seem to be enough CPU to run unRAID as a virtual machine. Have others used the same hardware for virtual machines? Try asking in the User customizations forum.
June 4, 201313 yr Author PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 1495 admin 20 0 16700 4532 3392 R 53 0.2 1:06.70 smbd 1223 root 20 0 47780 1476 572 S 49 0.1 1:57.65 shfs shfs is the user share system. There does not seem to be enough CPU to run unRAID as a virtual machine. Have others used the same hardware for virtual machines? Try asking in the User customizations forum. Really strange, so what make copying direct to cache drive is faster. I just tested on FTP clients as well - same issue as well. Any other suggestion ideas what it might be? I wil try asking in the User customizations forum.
June 4, 201313 yr Really strange, so what make copying direct to cache drive is faster.Less work for the CPU. User shares involve another layer of logic that requires more CPU time.
June 5, 201313 yr Really strange, so what make copying direct to cache drive is faster.Less work for the CPU. User shares involve another layer of logic that requires more CPU time. Correct. shfs is not involved in writes directly to cache. This is evident in the top output.
Archived
This topic is now archived and is closed to further replies.