Hi,
I'm seeing very high IO wait times on my unraid server. It seems to happen randomly, and I can't really track it down to a single thing. The samba process on Unraid starts becoming very unresponsive. You cannot "open" a folder in a Windows client, it will just sit as "unresponsive" for 30 second to 1 1/2 minutes. It does seem to happen more often when the unraid server has more files being copied in or out. It always happens when running preclear. Below is a snapshot from top, the unraid server is running preclear, but this happens even when I'm not running preclear.
top - 10:51:14 up 1 day, 19:12, 5 users, load average: 6.94, 6.27, 6.01
Tasks: 161 total, 1 running, 160 sleeping, 0 stopped, 0 zombie
Cpu(s): 7.4%us, 6.9%sy, 0.0%ni, 47.2%id, 37.9%wa, 0.0%hi, 0.6%si, 0.0%st
Mem: 6226524k total, 6102460k used, 124064k free, 460520k buffers
Swap: 0k total, 0k used, 0k free, 5349732k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
21132 root 20 0 2152 584 436 S 34 0.0 0:01.08 sum
21131 root 20 0 10508 8876 620 D 19 0.1 0:00.59 dd
3313 root 20 0 57900 9876 592 S 5 0.2 179:30.76 shfs
14332 xbox 20 0 20036 6968 5884 S 2 0.1 1:32.47 smbd
3238 root 20 0 0 0 0 S 1 0.0 49:45.87 unraidd
20735 root 20 0 3116 1284 744 S 1 0.0 0:00.07 rsync
20737 root 20 0 2972 616 228 S 1 0.0 0:00.06 rsync
21105 odenbach 20 0 20044 4632 3624 S 1 0.1 0:00.03 smbd
21125 xbox 20 0 20044 3760 2960 S 1 0.1 0:00.03 smbd
479 root 20 0 0 0 0 S 0 0.0 64:30.91 kswapd0
21124 root 20 0 2472 1020 756 R 0 0.0 0:00.03 top
1 root 20 0 828 284 240 S 0 0.0 0:07.00 init
2 root 20 0 0 0 0 S 0 0.0 0:00.06 kthreadd
3 root 20 0 0 0 0 S 0 0.0 1:24.87 ksoftirqd/0
5 root 0 -20 0 0 0 S 0 0.0 0:00.00 kworker/0:0H
7 root 0 -20 0 0 0 S 0 0.0 0:00.00 kworker/u:0H
8 root RT 0 0 0 0 S 0 0.0 0:01.48 migration/0
Its a fairly powerful box. I'm running unraid under esxi, I have dual hex core xeons, and 64GB of ram in the server. For the unraid VM i've given it 4 cores and 6GB ram. Running unraid 5.0.3 and passing through a M1015 controller in IT mode. THe high load is caused by IO wait, but I don't see anything specifically. Even if under high IO wait, I would still expect that samba would respond in a timely fashion, like in a few seconds, as opposed to several minutes. Right now when running preclear, my samba shares are completely unusable. I cant access them with a windows client. When not running preclear, it happens randomly at times. I probably have 70-100 open files at any given time on the server. Right now, I can't even do an "hdparm -tT /dev/xxx". It just sits and waits. When not in the degraded condition, I can copy files out of the server at wire speed, normally around 110 Megabytes per second or so. The degraded condition tends to come and go, but always happens when running preclear. I would expect samba to slow down when running preclear, but I don't think it should go completely unresponsive for minutes at a time.
Any thoughts on this?
Thanks.