Ultra slow read speeds 4.5MB/s

[rx8freehk]

Hello all, been running Unraid now for about 9 months and the speed is finally irratated me enought to try to figure out what the heck is going on.

 

Current Setup

 

unRAID Version: unRAID Server Plus, Version 5.0-rc10

Motherboard:         ASRock - N68C-S

Processor:         AMD AthlonTM II X3 450 - 3.2 GHz

Cache:                 L1 = 384 kB  L2 = 1536 kB 

Memory:                 2 GB - DIMM0 = 400 MHz  DIMM1 = 400 MHz 

Network:                 1000Mb/s - Full Duplex

 

sda Flash Drive  DT_Micro (7GB - 174MB Used)

sdd Cache Drive WDC_WD1001FALS-00E8B0 (Black) (1TB - 102GB Used)

sdh Parity Drive  Hitachi_HDS723020BLA642 (2TB)

sde HD1              Hitachi_HDS723020BLA642 (2TB - 1.88TB Used)

sdf HD2              Hitachi_HDS723020BLA642 (2TB - 1.93TB Used)

sdc HD3              TOSHIBA_DT01ACA200  (2TB - 1.21TB Used)

sdg HD4              WDC_WD20EARX-00PASB0 (Green) (2TB - 1.92TB Used)

sdb HD5              WDC_WD20EARX-00PASB0 (Green) (2TB - 1.55TB Used)

 

SD A,H,E,F are connected to MB

SD D,G are connected to Silicon Image SIL3132 (1)

SD B is connected to another Silicon Image SIL3132 (2)

 

Packages running -

Plex Media Server

SABnzbd+

Sickbeard

Couchpotato

Transmission

Mysql

 

Nothing Currently being downloads/processed/viewed/etc

 

Im pulling approx 4.3MB/s transfering 7GB .iso files from one of my unraid shares (not cached) and typically i max out around 12MB/s.

 

Any/all help is greatly appreciated!

syslog.txt

[Joe L.]

what do you see when you type:

ethtool eth0

and

ifconfig eth0

[rx8freehk]

 

ethtool eth0

 

Settings for eth0:

        Supported ports: [ TP ]

        Supported link modes:  10baseT/Half 10baseT/Full

                                100baseT/Half 100baseT/Full

                                1000baseT/Full

        Supports auto-negotiation: Yes

        Advertised link modes:  10baseT/Half 10baseT/Full

                                100baseT/Half 100baseT/Full

                                1000baseT/Full

        Advertised pause frame use: Symmetric Receive-only

        Advertised auto-negotiation: Yes

        Speed: 1000Mb/s

        Duplex: Full

        Port: Twisted Pair

        PHYAD: 0

        Transceiver: internal

        Auto-negotiation: on

        MDI-X: Unknown

        Supports Wake-on: pumbg

        Wake-on: g

        Current message level: 0x00000033 (51)

        Link detected: yes

 

 

ipconfig eth0

 

eth0      Link encap:Ethernet  HWaddr 00:0a:cd:1b:57:79

          inet addr:192.168.1.111  Bcast:192.168.1.255  Mask:255.255.255.0

          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1

          RX packets:116000805 errors:0 dropped:76 overruns:0 frame:0

          TX packets:104131638 errors:0 dropped:0 overruns:0 carrier:0

          collisions:0 txqueuelen:1000

          RX bytes:2523439851 (2.3 GiB)  TX bytes:712048646 (679.0 MiB)

          Interrupt:18 Base address:0xec00

 

 

 

[RobJ]

I have a number of comments, hopefully helpful:

 

1. The "ipconfig eth0" command showed 76 dropped packets, and while that may not seem like many, anything non-zero seems to have an outsized impact on network quality and speed.  Your syslog also showed the network disconnected and reconnected several times.  This all translates into slower network speed.  I would check for network interference, such as network cables running too close to power cords.  Possibly a replacement network cable may also help.

 

2. Disks 1, 2 and 4 are almost full.  One of the weaknesses of the Reiser file system is that when you get too close to filling up a drive, adding new files becomes extremely slow.  If you are adding to User Shares, I would adjust the Minimum Free to something like 50GB or more, or add another drive to the share, or replace drive with larger.

 

3. You have only 1GB only 2GB of RAM, and a LOT of addons, and some are probably heavy RAM users, like SQLite and Apache and SimpleFeatures and probably others.  This may mean poor disk buffering, which will affect all disk I/O.  I wonder if there is some disk thrashing going on too.  More RAM might help.

 

4. You are using 2 SiI3132 addon cards, which I like too, but are older technology now, and cap speeds to UDMA/100.  That means it is very hard to get full speed I/O to your newer drives.

 

5. From the speeds you report, it does not sound like you are writing to the Cache drive.  Are you sure your User Shares are configured to use the Cache disk?

 

6. I recommend upgrading to RC11.  You are almost there with RC10, but there *might* be something fixed in RC11 that *might* help you.

 

7. Possibly minor but odd, Disk 2 (sdf) connects at 1.5Gbps, whereas its identical twin and all other drives connect at 3.0 Gbps.  I see no apparent reason for it.  I don't think it will noticeably affect the I/O speed any, but is odd.

 

8. On Mar 2 at 21:38:05, there is an apparent attempt to shutdown for reboot, but unsuccessful.  No idea what happened there, or the ramifications of that attempt.

 

9. On Mar 3 at 17:32:12, there was a segfault in Transmission.  About 30 minutes later, there is an apparent restart of Transmission, with possible configuration warnings.  I would not trust the operation of a subsystem after a segfault until after a reboot.

 

There are no disk errors, so that's one good thing!

 

Edit: I misread the amount of RAM, the way it's divided up, so you are using 2GB, minus 256MB for the onboard video.  You might check the BIOS video settings, to see if you can reduce that 256MB as low as possible.  Your text mode consoles don't need hardly any.

[rx8freehk]

Haven't had time to dive into this just yet, I did order 8GB (2x 4GB) of  RAM and will tackle the other suggestions this weekend. Really appreciate the time you took for such an in depth analysis of the logs!

[madburg]

I have a number of comments, hopefully helpful:

 

1. The "ifconfig eth0" command showed 76 dropped packets, and while that may not seem like many, anything non-zero seems to have an outsized impact on network quality and speed.

 

Hey RobJ, for what its worth, I have been monitoring this for a while on my setup (also on my layer 2 and 3 switches which show no issue) and noticed the following whenever unRAID is rebooted several dropped packets are logged. As an example my uptime is 15days and I have 37 (received) dropped  packets (all other stats are zero'ed under ifconfig), these are the initial dropped packets when unRAID first comes on. To be clear it not always the same value (ex. 37 packets).

 

So his 76 may not be out of place.