January 28, 201016 yr Hi looking for some help, I'm out of idea's. Fairly long term unraid user. I'm currently plagued by stuttering. i've removed all add ons and am running unraid 4.5 stock. Playing back HD files record by sage is brutal. If I pause the file for a few seconds, it will play stutter free for a few seconds. doing a top, I see that shfs is taking 100% of 1 cpu. This seems to me excessive. The box is a 2 core amd, 1 Gig of ram. all drives are attached to the mobo ports - 6 sata, 2 ide. The web front end and telnet-ting to the box are quick. Copying a file from the box is also slow, utilizing only 12.5 % of a 100 mbit network connection. the unraid server is attached to a gig switch at 1000mbits, and the client pc (xp) is connected to the same switch a 100mbits. Any Idea's ? top - 20:56:11 up 2 days, 23:17, 1 user, load average: 0.00, 0.00, 0.01 Mem: 968556k total, 957620k used, 10936k free, 80100k buffers Swap: 0k total, 0k used, 0k free, 685276k cached Unknown command - try 'h' for help PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 1448 root 20 0 127m 11m 624 S 100 1.2 717:11.32 shfs 1319 root 20 0 13656 3400 2640 S 0 0.4 61:28.10 smbd 5552 root 20 0 20728 10m 3408 S 0 1.1 13:56.89 smbd 1310 root 20 0 1880 656 548 S 0 0.1 12:46.23 crond 1262 root 20 0 1728 232 156 S 0 0.0 8:48.54 dhcpcd 1455 root 20 0 26988 1196 544 S 0 0.1 5:18.37 shfs 1375 root 15 -5 0 0 0 S 0 0.0 2:09.71 unraidd 259 root 15 -5 0 0 0 S 0 0.0 1:02.33 kswapd0 5564 root 20 0 14340 4360 3168 S 0 0.5 0:17.56 smbd 5120 root 20 0 0 0 0 S 0 0.0 0:04.33 pdflush 1 root 20 0 776 304 264 S 0 0.0 0:02.64 init 1317 root 20 0 8540 1908 1304 S 0 0.2 0:01.31 nmbd 1330 root 20 0 53904 1576 1184 S 0 0.2 0:00.79 emhttp 1426 root 15 -5 0 0 0 S 0 0.0 0:00.75 reiserfs/1 581 root 16 -4 1840 668 448 S 0 0.1 0:00.59 udevd 1425 root 15 -5 0 0 0 S 0 0.0 0:00.41 reiserfs/0 1189 root 20 0 1752 592 504 S 0 0.1 0:00.32 syslogd 1232 root 20 0 1732 464 392 S 0 0.0 0:00.11 ifplugd 108 root 15 -5 0 0 0 S 0 0.0 0:00.07 kblockd/1 I've attached the syslog. syslog.txt
January 28, 201016 yr I'm far from an expert here but you seem to have duplicate files and the cache drive/mover script is involved. Have you tried to tell sage to write directly to \\tower\diskX\sage instead of \\tower\sage ? Gog
January 28, 201016 yr You think loosing a lot of packets could pin the CPU to 100% ? Yup, there are situations where network problems can cause high CPU.... particularly if your NIC is not offloading work from the CPU. And there is no guarantee that the two symptoms are caused by the same thing. Connection resets may being passed down the stack with Samba, which is causing all the extra work in shfs.
January 28, 201016 yr Author Thanks for the replies. The stuttering happened even before I started using the cache drive. In fact that is why I added it. I'll check on the switch and cables, I believe the cables are cat 5e. Is there anything I can run on the server or client to get stats on packet loss?
January 28, 201016 yr Thanks for the replies. The stuttering happened even before I started using the cache drive. In fact that is why I added it. It would only help in writing speed when writing files to the array... I doubt if it would help in any stuttering issue. I'll check on the switch and cables, I believe the cables are cat 5e. Did you make them? or were the RJ45 connectors factory crimped? The reason I ask is there are two different "standards" for wiring the pairs of cat5e to the connectors. One is for telephone use, the other for LAN use. The "crimping tool set" I purchased had a nice color-coded guide in the carrying case lid... for telephone use. It made no mention of the correct wiring for LAN use. If you did not know any better, it would be easy to wire the connectors wrong when constructing a cable for LAN use. Both wiring "standards" would give continuity from pin to pin on the ends, but only one would work well at any length past a few feet on a LAN. (if that) Is there anything I can run on the server or client to get stats on packet loss? Yes, logged in on the unRAID server, to see statistics type ifconfig eth0 to see the network speed and how it has negotiated a connection type: ethtool eth0
January 29, 201016 yr Author Well, Network cables are store bought, both cat 5 - no e, the switch is a trendnet teg 680g. however this is the result from ifconfig and ethtool to me it looks like no dropped packets. interestingly I just did another copy - ~90 utilization.. top at that point was showing <10% to smbd Something else must be going on. root@Tower:~# ifconfig eth0 eth0 Link encap:Ethernet HWaddr 00:1f:d0:9f:00:a2 inet addr:192.168.1.4 Bcast:192.168.1.255 Mask:255.255.255.0 UP BROADCAST NOTRAILERS RUNNING MULTICAST MTU:1500 Metric:1 RX packets:114978978 errors:0 dropped:0 overruns:0 frame:0 TX packets:119715771 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:1833707904 (1.7 GiB) TX bytes:4084590972 (3.8 GiB) Interrupt:26 Base address:0xa000 root@Tower:~# ethtool eth0 Settings for eth0: Supported ports: [ TP MII ] Supported link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Half 1000baseT/Full Supports auto-negotiation: Yes Advertised link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Half 1000baseT/Full Advertised auto-negotiation: Yes Speed: 1000Mb/s Duplex: Full Port: MII PHYAD: 0 Transceiver: internal Auto-negotiation: on Supports Wake-on: pumbg Wake-on: g Current message level: 0x00000033 (51) Link detected: yes root@Tower:~#
January 29, 201016 yr Playing back HD files record by sage is brutal. Since you're using SageTV to play HD content, then you probably shouldn't bother with user shares. Set up Sage's import directories directly to your unRAID's disk shares. SageTV is smart enough to give you an aggregate view of all its import directories. This way you'll still achieve the same result as with user shares, but without the extra overhead of the user shares.
January 29, 201016 yr I use sage, user shares and a HRHD for HD recording. I had no problem playing a HD file with 4.4.2 but the array had to be quiet. Not completely idle, torrent were still going for example but no massive file move. Simultaneous recording and playing of a HD stream was not guaranteed to be pain free. I have not had any problems playing HD with 4.5 so far. I know I've done it while a SD stream was recorded and I *think* I also did it with one Sd and one HD recording. I no longer have issues playing media while moving files on the array. And I *love* that fact that user shares allow me to move sage recording from one drive to the other without having to worry about sage loosing my recordings. Gog
January 30, 201016 yr Author That's good to know that it can work. Another data point. Was watching a movie, it started to stutter. Checked top on the unraid server, shfs was > 90%. I used the unraid web site to take the array off-line, and then started it again. After that it was fine. Is there anyway to see what shfs is doing when it is so busy?
January 30, 201016 yr Armbrust, as I mentioned above, shfs normally brings a signifficant overhead. It's really not needed for SageTV.
January 30, 201016 yr Author Purko, yeah I know, and maybe I will resort to doing it that way, but I thought it should work using user shares. I'll continue to try and isolate the cause.... Thanks for looking at it.
February 10, 201016 yr Author Just to follow up, The main issue appears to have been using a user share as the target for utorrent downloads from a windows pc. I've switched to using a disk share for this purpose, and haven't seen the high cpu usage since. Stuttering seems to be largely eliminated as well.
February 11, 201016 yr If you want to store the completed torrents in unRaid, have you tried downloading to your pc, then having uTorrent automatically move the completed files to the server? That's how I have it, and no issues with stuttering or high cpu usage...
February 11, 201016 yr If you want to store the completed torrents in unRaid, have you tried downloading to your pc, then having uTorrent automatically move the completed files to the server? Right! Or, better yet, let the unRAID server do all the torrenting, but do it on a disk that's outside the protected array. And then have it automatically move the completed files to the protected array.
Archived
This topic is now archived and is closed to further replies.