itsrumsey

Members
  • Posts

    46
  • Joined

  • Last visited

Converted

  • Gender
    Male
  • Location
    Dallas, TX

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

itsrumsey's Achievements

Rookie

Rookie (2/14)

0

Reputation

  1. Question for those using ESXi. I understand the free version only support 8 vCPUs. If I am running a dual processor configuration, with 8 cores and 16 threads each, for a total of 32 threads and 16 cores... can I utilize all of these in one VM on the free edition?
  2. I have been paying $125 a month for Charter Business internet for over 3 years now for a paltry 100/7 connection and due to 4 outages in the last month, have decided I have had it. I want to move my server to a colo and have been shopping around for deals. The problem is the only decent ones I can find are 1U-4U and my server itself is in a 4U chassis, that means no room to bring my own firewall. I've been thinking it through and best as I can figure there are two options: run ESXi or similar on the metal and virtualize my unRAID installation, then run pfSense in another VM run pfSense in a VM under unRAID, using another NIC and passthrough If I choose the latter I feel like I am in for quite the challenge, as you could categorize my linux/unix confidence as firmly in "beginner". Conversely, if I go for option 1 I am not sure how difficult it would be to virtualize my existing unRAID configuration. Maybe someone who has done it (or has knowledge how) could comment, but the way I see running pfSense within unRAID I'd need to utilize dual NIC on my motherboard as well as purchase another dual NIC card to passthrough. We'll call the motherboard NICs eth0 / eth1 and the addin card pci1 / pci2. I figure pci1 would accept the publicly open network connection from the colo, pfSense would act as the router / firewall, there would be a crossover cable from pci2 to eth0, eth0 would be in the primary incoming NIC for unRAID and would obtain IP lease etc from pfSense VM, and eth1 would be reserved for IPMI management. Does this sound about right? Is there any way it could be simplified further? If it is mostly trivial to virtualize my existing unRAID configuration in ESXi, that may be the easier choice since I can do all of this with virtual NICs and won't be researching KVM / linux network configuration for days.
  3. https://www.supermicro.com/products/motherboard/xeon/c600/x9dr3-ln4f_.cfm Is anyone using this board with unRAID? I am right on the brink of ordering it, but I can't find any solid information on whether the SAS controller is supported. I found these 2 previous threads which didn't inspire confidence: I'd have to blow $350 on eBay only to find out I can't use the board in my server.
  4. That would work except the permissions on the folders say they belong to unix\1000 (the UID Emby docker runs under, I guess) and I can't do it through SMB. I haven't tried logging in to the share as root but that might work.
  5. Hey guys, simple question here. Can someone share the command syntax for searching a directory (and all sub directories) for a specific folder name, and then deleting those folders and all their contents? Context: I want to search /mnt/user/movies for folders named "extrathumbs" and delete both the folder (there will be thousands) and their contents.
  6. Is anyone else having issues with this docker? I have been using it ~2 years and now it is stuck on 3.2.5, Emby keeps spamming me that 3.2.8 is available. Docker shows up to date in my docker manager: Tried both of these to no avail as well.
  7. In the end I pulled out the server and ripped it apart drive by drive to find the culprit. It was the cache drive. For anyone else experiencing this issues here are some clues in the log: Dec 27 23:30:48 Tower emhttp: Kingston_DataTraveler_2.0_50E5495131BBBF50892D0641-0:0 (sda) 15364384 Dec 27 23:30:48 Tower emhttp: ST3000DM001-1CH166_W1F22GP5 (sdb) 2930266532 Dec 27 23:30:48 Tower emhttp: TOSHIBA_MD04ACA500_55TAK2CWFS9A (sdc) 4883770532 Dec 27 23:30:49 Tower emhttp: ST3000DM001-1CH166_W1F24BS0 (sdd) 2930266532 Dec 27 23:30:49 Tower emhttp: TOSHIBA_MD04ACA500_6566KAFQFS9A (sde) 4883770532 Dec 27 23:30:49 Tower emhttp: ST5000DM000-1FK178_W4J01V9W (sdf) 4883770532 Dec 27 23:30:49 Tower emhttp: TOSHIBA_MD04ACA500_55TAK2CVFS9A (sdg) 4883770532 Dec 27 23:30:49 Tower emhttp: TOSHIBA_MD04ACA500_55TAK2CUFS9A (sdh) 4883770532 Dec 27 23:30:49 Tower kernel: mdcmd (1): import 0 sdc 4883770532 0 TOSHIBA_MD04ACA500_55TAK2CWFS9A Dec 27 23:30:49 Tower kernel: md: import disk0: (sdc) TOSHIBA_MD04ACA500_55TAK2CWFS9A size: 4883770532 Dec 27 23:30:49 Tower kernel: mdcmd (2): import 1 sdh 4883770532 0 TOSHIBA_MD04ACA500_55TAK2CUFS9A Dec 27 23:30:49 Tower kernel: md: import disk1: (sdh) TOSHIBA_MD04ACA500_55TAK2CUFS9A size: 4883770532 Dec 27 23:30:49 Tower kernel: mdcmd (3): import 2 sdg 4883770532 0 TOSHIBA_MD04ACA500_55TAK2CVFS9A Dec 27 23:30:49 Tower kernel: md: import disk2: (sdg) TOSHIBA_MD04ACA500_55TAK2CVFS9A size: 4883770532 Dec 27 23:30:49 Tower kernel: mdcmd (4): import 3 sdb 2930266532 0 ST3000DM001-1CH166_W1F22GP5 Dec 27 23:30:49 Tower kernel: md: import disk3: (sdb) ST3000DM001-1CH166_W1F22GP5 size: 2930266532 Dec 27 23:30:51 Tower kernel: mdcmd (5): import 4 sdf 4883770532 0 ST5000DM000-1FK178_W4J01V9W Dec 27 23:30:51 Tower kernel: md: import disk4: (sdf) ST5000DM000-1FK178_W4J01V9W size: 4883770532 Dec 27 23:30:51 Tower kernel: mdcmd (6): import 5 sde 4883770532 0 TOSHIBA_MD04ACA500_6566KAFQFS9A Dec 27 23:30:51 Tower kernel: md: import disk5: (sde) TOSHIBA_MD04ACA500_6566KAFQFS9A size: 4883770532 Dec 27 23:30:51 Tower kernel: mdcmd (7): import 6 Dec 27 23:30:51 Tower kernel: mdcmd (: import 7 Dec 27 23:30:51 Tower kernel: mdcmd (9): import 8 Dec 27 23:30:51 Tower kernel: mdcmd (10): import 9 Dec 27 23:30:51 Tower kernel: mdcmd (11): import 10 Dec 27 23:30:51 Tower kernel: mdcmd (12): import 11 Dec 27 23:30:51 Tower kernel: mdcmd (13): import 12 Dec 27 23:30:51 Tower kernel: mdcmd (14): import 13 Dec 27 23:30:51 Tower kernel: mdcmd (15): import 14 Dec 27 23:30:51 Tower kernel: mdcmd (16): import 15 Dec 27 23:30:51 Tower kernel: mdcmd (17): import 16 Dec 27 23:30:51 Tower kernel: mdcmd (18): import 17 Dec 27 23:30:51 Tower kernel: mdcmd (19): import 18 Dec 27 23:30:51 Tower kernel: mdcmd (20): import 19 Dec 27 23:30:51 Tower kernel: mdcmd (21): import 20 Dec 27 23:30:51 Tower kernel: mdcmd (22): import 21 Dec 27 23:30:51 Tower kernel: mdcmd (23): import 22 Dec 27 23:30:51 Tower kernel: mdcmd (24): import 23 Dec 27 23:30:51 Tower kernel: mdcmd (25): import 24 Dec 27 23:30:51 Tower kernel: mdcmd (26): import 25 Dec 27 23:30:51 Tower kernel: mdcmd (27): import 26 Dec 27 23:30:51 Tower kernel: mdcmd (28): import 27 Dec 27 23:30:51 Tower kernel: mdcmd (29): import 28 Dec 27 23:30:51 Tower kernel: mdcmd (30): import 29 Dec 27 23:30:51 Tower kernel: md: import_slot: 29 empty Excluding the boot drive, drives B, C, D, E, F, G, H are listed at the beginning, then only C, H, G, B, F, E are imported... You may notice D is missing. This drive was the culprit. I tried another SATA port, another cable, but in the end the drive seemed to be the issue. Replace the drive and all is fine.
  8. I am getting a 30-60 second delay every time I switch tabs on the Unraid dashboard which seem to be the result of a hardware error. The problem is, I don't know how to identify which disk is the issue. It shows up plainly in the logs: Dec 27 23:30:48 Tower emhttp: shcmd (90): rmmod md-mod |& logger Dec 27 23:30:48 Tower kernel: md: unRAID driver removed Dec 27 23:30:48 Tower emhttp: shcmd (91): modprobe md-mod super=/boot/config/super.dat |& logger Dec 27 23:30:48 Tower kernel: md: unRAID driver 2.6.6 installed Dec 27 23:30:48 Tower emhttp: Plus key detected, GUID: XXXX-XXXX-31BB-BF50892D0641 FILE: /boot/config/Plus.key Dec 27 23:30:48 Tower emhttp: Device inventory: Dec 27 23:30:48 Tower emhttp: shcmd (92): udevadm settle Dec 27 23:30:48 Tower emhttp: Kingston_DataTraveler_2.0_50E5495131BBBF50892D0641-0:0 (sda) 15364384 Dec 27 23:30:48 Tower emhttp: ST3000DM001-1CH166_W1F22GP5 (sdb) 2930266532 Dec 27 23:30:48 Tower emhttp: TOSHIBA_MD04ACA500_55TAK2CWFS9A (sdc) 4883770532 Dec 27 23:30:49 Tower emhttp: ST3000DM001-1CH166_W1F24BS0 (sdd) 2930266532 Dec 27 23:30:49 Tower emhttp: TOSHIBA_MD04ACA500_6566KAFQFS9A (sde) 4883770532 Dec 27 23:30:49 Tower emhttp: ST5000DM000-1FK178_W4J01V9W (sdf) 4883770532 Dec 27 23:30:49 Tower emhttp: TOSHIBA_MD04ACA500_55TAK2CVFS9A (sdg) 4883770532 Dec 27 23:30:49 Tower emhttp: TOSHIBA_MD04ACA500_55TAK2CUFS9A (sdh) 4883770532 Dec 27 23:30:49 Tower kernel: mdcmd (1): import 0 sdc 4883770532 0 TOSHIBA_MD04ACA500_55TAK2CWFS9A Dec 27 23:30:49 Tower kernel: md: import disk0: (sdc) TOSHIBA_MD04ACA500_55TAK2CWFS9A size: 4883770532 Dec 27 23:30:49 Tower kernel: mdcmd (2): import 1 sdh 4883770532 0 TOSHIBA_MD04ACA500_55TAK2CUFS9A Dec 27 23:30:49 Tower kernel: md: import disk1: (sdh) TOSHIBA_MD04ACA500_55TAK2CUFS9A size: 4883770532 Dec 27 23:30:49 Tower kernel: mdcmd (3): import 2 sdg 4883770532 0 TOSHIBA_MD04ACA500_55TAK2CVFS9A Dec 27 23:30:49 Tower kernel: md: import disk2: (sdg) TOSHIBA_MD04ACA500_55TAK2CVFS9A size: 4883770532 Dec 27 23:30:49 Tower kernel: mdcmd (4): import 3 sdb 2930266532 0 ST3000DM001-1CH166_W1F22GP5 Dec 27 23:30:49 Tower kernel: md: import disk3: (sdb) ST3000DM001-1CH166_W1F22GP5 size: 2930266532 Dec 27 23:30:51 Tower kernel: mdcmd (5): import 4 sdf 4883770532 0 ST5000DM000-1FK178_W4J01V9W Dec 27 23:30:51 Tower kernel: md: import disk4: (sdf) ST5000DM000-1FK178_W4J01V9W size: 4883770532 Dec 27 23:30:51 Tower kernel: mdcmd (6): import 5 sde 4883770532 0 TOSHIBA_MD04ACA500_6566KAFQFS9A Dec 27 23:30:51 Tower kernel: md: import disk5: (sde) TOSHIBA_MD04ACA500_6566KAFQFS9A size: 4883770532 Dec 27 23:30:51 Tower kernel: mdcmd (7): import 6 Dec 27 23:30:51 Tower kernel: mdcmd (: import 7 Dec 27 23:30:51 Tower kernel: mdcmd (9): import 8 Dec 27 23:30:51 Tower kernel: mdcmd (10): import 9 Dec 27 23:30:51 Tower kernel: mdcmd (11): import 10 Dec 27 23:30:51 Tower kernel: mdcmd (12): import 11 Dec 27 23:30:51 Tower kernel: mdcmd (13): import 12 Dec 27 23:30:51 Tower kernel: mdcmd (14): import 13 Dec 27 23:30:51 Tower kernel: mdcmd (15): import 14 Dec 27 23:30:51 Tower kernel: mdcmd (16): import 15 Dec 27 23:30:51 Tower kernel: mdcmd (17): import 16 Dec 27 23:30:51 Tower kernel: mdcmd (18): import 17 Dec 27 23:30:51 Tower kernel: mdcmd (19): import 18 Dec 27 23:30:51 Tower kernel: mdcmd (20): import 19 Dec 27 23:30:51 Tower kernel: mdcmd (21): import 20 Dec 27 23:30:51 Tower kernel: mdcmd (22): import 21 Dec 27 23:30:51 Tower kernel: mdcmd (23): import 22 Dec 27 23:30:51 Tower kernel: mdcmd (24): import 23 Dec 27 23:30:51 Tower kernel: mdcmd (25): import 24 Dec 27 23:30:51 Tower kernel: mdcmd (26): import 25 Dec 27 23:30:51 Tower kernel: mdcmd (27): import 26 Dec 27 23:30:51 Tower kernel: mdcmd (28): import 27 Dec 27 23:30:51 Tower kernel: mdcmd (29): import 28 Dec 27 23:30:51 Tower kernel: mdcmd (30): import 29 Dec 27 23:30:51 Tower kernel: md: import_slot: 29 empty Dec 27 23:30:51 Tower emhttp: import 30 cache device: sdd Dec 27 23:30:51 Tower emhttp: import flash device: sda Dec 27 23:31:12 Tower kernel: ata3.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen Dec 27 23:31:12 Tower kernel: ata3.00: failed command: SMART Dec 27 23:31:12 Tower kernel: ata3.00: cmd b0/d0:01:00:4f:c2/00:00:00:00:00/00 tag 14 pio 512 in Dec 27 23:31:12 Tower kernel: res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout) Dec 27 23:31:12 Tower kernel: ata3.00: status: { DRDY } Dec 27 23:31:12 Tower kernel: ata3: hard resetting link Dec 27 23:31:17 Tower kernel: ata3: link is slow to respond, please be patient (ready=0) Dec 27 23:31:20 Tower kernel: ata3: SATA link up 6.0 Gbps (SStatus 133 SControl 300) Dec 27 23:31:20 Tower kernel: ata3.00: configured for UDMA/133 Dec 27 23:31:20 Tower kernel: ata3: EH complete What in the heck is going on here?
  9. Thank you god for being the last post in this thread.
  10. Getting RDP-Calibre up and running was a snap, but now that it is I feel uneasy about exposing the library through my firewall. Is there any way to set up password protection? I see through the GUI you can configure more options for sharing including user and password, but this stuff seems to be irrelevant as it appears the server share portion of the docker is running detached from this server process?
  11. Do you think the limit on bandwidth there is the device or the bus? That is good for one device but I don't know about splitting 4 over that same speed.
  12. One more really dumb question I don't want to research too hard. What is the minimum PCI-Express speed to reach 6GB/s? I see PCI-Express 2.0 cards at 1x advertising 6GB/s, is this legit? Say I want a 4 port card, which speed slot could actually handle the bandwidth? 4x? 16x? Also, any comment on the http://www.newegg.com/Product/Product.aspx?Item=N82E16813157500R board I am considering?
  13. Wise man. Makes perfect sense and I feel like a fool now for even asking. Well, that's what 10 beers and then deciding to rebuild your server does for you Going to replace all the hardware as well so I hope I don't run in to any issues. I think unRAID activation is tied to the USB key and not some other hardware identifiers, correct me if I'm wrong?
  14. Hey guys, here's what I'm looking to do. I know the "official" procedure for upgrading a cache drive. I'm not super familiar with linux in general. I don't want to back up my perfectly good cache drive to my array and then copy it back to a new cache drive. What I want to know how to do is insert a new disk (new cache drive), format it, mount it, and copy the current cache drives contents to it... Then power down, remove my cache drive, and reassign this new disk as the cache drive. I can't think of any reason why formatting a drive or mounting a drive should be impossible in unRAID without adding it to the array. I believe the unassigned devices addon does exactly this - mounts and shares drives outside the array. Can someone give me some straightforward tips on accomplishing the obvious?
  15. Hoping binhex or someone can help out here. I've been running this docker forever and not had any issues with it. Today I wanted to modify which cores it ran on so I stopped the docker (after shutting down the daemon through web console), clicked edit, changed the cores it is running on, and clicked apply. That cause the docker to pull some image updates and after it completed, Deluge no longer launches correctly and web console never loads. Here is a log of the docker when it starts: usermod: no changes [info] Env var PUID defined as 99 [info] Env var PGID defined as 100 [info] Permissions already set for /config [info] Starting Supervisor... 2016-03-31 23:24:42,488 CRIT Set uid to user 0 2016-03-31 23:24:42,495 WARN Included extra file "/etc/supervisor/conf.d/delugevpn.conf" during parsing 2016-03-31 23:24:42,501 INFO supervisord started with pid 13 2016-03-31 23:24:43,504 INFO spawned: 'privoxy' with pid 16 2016-03-31 23:24:43,505 INFO spawned: 'start' with pid 17 2016-03-31 23:24:43,507 INFO spawned: 'webui' with pid 18 2016-03-31 23:24:43,508 INFO spawned: 'deluge' with pid 19 2016-03-31 23:24:43,524 DEBG 'privoxy' stdout output: [info] VPN is enabled, checking VPN tunnel local ip is valid 2016-03-31 23:24:43,524 INFO success: privoxy entered RUNNING state, process has stayed up for > than 0 seconds (startsecs) 2016-03-31 23:24:43,524 INFO success: start entered RUNNING state, process has stayed up for > than 0 seconds (startsecs) 2016-03-31 23:24:43,531 DEBG 'privoxy' stdout output: [info] checking VPN tunnel local ip is valid... 2016-03-31 23:24:43,536 DEBG 'deluge' stdout output: [info] VPN enabled, configuring Deluge... 2016-03-31 23:24:43,537 DEBG 'deluge' stdout output: [info] checking VPN tunnel local ip is valid... 2016-03-31 23:24:43,541 DEBG 'start' stdout output: [info] VPN is enabled, beginning configuration of VPN 2016-03-31 23:24:43,570 DEBG 'start' stdout output: [info] VPN provider defined as pia [info] VPN config file (ovpn extension) is located at /config/openvpn/openvpn.ovpn 2016-03-31 23:24:43,573 DEBG 'start' stdout output: [info] Env vars not defined for remote host, port and protocol, will parse existing entries from ovpn file... 2016-03-31 23:24:43,609 DEBG 'start' stdout output: [info] VPN provider remote gateway defined as ca-toronto.privateinternetaccess.com [info] VPN provider remote port defined as 1194 [info] VPN provider remote protocol defined as udp 2016-03-31 23:24:43,616 DEBG 'start' stdout output: [info] VPN provider username defined as pXXXXXXX 2016-03-31 23:24:43,620 DEBG 'start' stdout output: [info] VPN provider password defined as XXXXXXXXXX 2016-03-31 23:24:43,635 DEBG 'start' stdout output: [info] Default route for container is 172.17.42.1 2016-03-31 23:24:43,654 DEBG 'start' stdout output: [crit] LAN network not defined, please specify via env variable LAN_NETWORK 2016-03-31 23:24:43,655 DEBG fd 9 closed, stopped monitoring (stdout)> 2016-03-31 23:24:43,655 DEBG fd 14 closed, stopped monitoring (stderr)> 2016-03-31 23:24:43,655 INFO exited: start (exit status 1; not expected) 2016-03-31 23:24:43,655 DEBG received SIGCLD indicating a child quit 2016-03-31 23:24:44,656 INFO success: webui entered RUNNING state, process has stayed up for > than 1 seconds (startsecs) 2016-03-31 23:24:44,656 INFO success: deluge entered RUNNING state, process has stayed up for > than 1 seconds (startsecs) Nothing else ever follows that and the console never loads. edit: Ok so it looks like the environment variable LAN_RANGE was replaced with LAN_NETWORK at some point and that is where the problem seems to be. I tried simply renaming LAN_RANGE to LAN_NETWORK but then I get the following errors, so I assume the syntax was changed as well: 2016-03-31 23:35:34,071 DEBG 'start' stdout output: [info] Default route for container is 172.17.42.1 2016-03-31 23:35:34,081 DEBG 'start' stdout output: [info] [u][b]Adding 192.168.0.1-192.168.0.254 as route via docker eth0[/b][/u] 2016-03-31 23:35:34,082 DEBG 'start' stderr output: [u][b]Error: an inet prefix is expected rather than "192.168.0.1-192.168.0.254".[/b][/u] Now it seems to at least launch and download, but the above message has me concerned it is messing up the routing table and potentially not using VPN. Here's an extended excerpt showing some concerning messages: 2016-03-31 23:35:33,259 DEBG 'start' stdout output: [info] Env vars not defined for remote host, port and protocol, will parse existing entries from ovpn file... 2016-03-31 23:35:34,045 DEBG 'start' stdout output: [info] VPN provider remote gateway defined as ca-toronto.privateinternetaccess.com [info] VPN provider remote port defined as 1194 [info] VPN provider remote protocol defined as udp 2016-03-31 23:35:34,053 DEBG 'start' stdout output: [info] VPN provider username defined as pXXXXXXX 2016-03-31 23:35:34,057 DEBG 'start' stdout output: [info] VPN provider password defined as XXXXXXXXXX 2016-03-31 23:35:34,071 DEBG 'start' stdout output: [info] Default route for container is 172.17.42.1 2016-03-31 23:35:34,081 DEBG 'start' stdout output: [u][b][info] Adding 192.168.0.1-192.168.0.254 as route via docker eth[/b][/u]0 2016-03-31 23:35:34,082 DEBG 'start' stderr output: [u][b]Error: an inet prefix is expected rather than "192.168.0.1-192.168.0.254"[/b][/u]. 2016-03-31 23:35:34,082 DEBG 'start' stdout output: [info] ip route defined as follows... -------------------- 2016-03-31 23:35:34,083 DEBG 'start' stdout output: default via 172.17.42.1 dev eth0 2016-03-31 23:35:34,083 DEBG 'start' stdout output: 172.17.0.0/16 dev eth0 proto kernel scope link src 172.17.0.20 -------------------- 2016-03-31 23:35:34,090 DEBG 'start' stdout output: [info] iptable_mangle support detected, adding fwmark for tables 2016-03-31 23:35:35,092 INFO success: webui entered RUNNING state, process has stayed up for > than 1 seconds (startsecs) 2016-03-31 23:35:35,092 INFO success: deluge entered RUNNING state, process has stayed up for > than 1 seconds (startsecs) 2016-03-31 23:36:14,188 DEBG 'start' stderr output: [u][b]iptables v1.4.21: host/network `192.168.0.1-192.168.0.254' not found Try `iptables -h' or 'iptables --help' for more information. 2016-03-31 23:36:14,205 DEBG 'start' stderr output: iptables v1.4.21: host/network `192.168.0.1-192.168.0.254' not found Try `iptables -h' or 'iptables --help' for more information.[/b][/u] 2016-03-31 23:36:14,209 DEBG 'start' stdout output: [info] iptables defined as follows... -------------------- 2016-03-31 23:36:14,210 DEBG 'start' stdout output: -P INPUT DROP -P FORWARD ACCEPT -P OUTPUT DROP -A INPUT -i tun0 -j ACCEPT -A INPUT -s 172.17.0.0/16 -d 172.17.0.0/16 -j ACCEPT -A INPUT -i eth0 -p udp -m udp --sport 1194 -j ACCEPT -A INPUT -i eth0 -p tcp -m tcp --dport 8112 -j ACCEPT -A INPUT -i eth0 -p tcp -m tcp --sport 8112 -j ACCEPT -A INPUT -p udp -m udp --sport 53 -j ACCEPT -A INPUT -p icmp -m icmp --icmp-type 0 -j ACCEPT -A INPUT -i lo -j ACCEPT -A OUTPUT -o tun0 -j ACCEPT -A OUTPUT -s 172.17.0.0/16 -d 172.17.0.0/16 -j ACCEPT -A OUTPUT -o eth0 -p udp -m udp --dport 1194 -j ACCEPT -A OUTPUT -o eth0 -p tcp -m tcp --dport 8112 -j ACCEPT -A OUTPUT -o eth0 -p tcp -m tcp --sport 8112 -j ACCEPT -A OUTPUT -p udp -m udp --dport 53 -j ACCEPT -A OUTPUT -p icmp -m icmp --icmp-type 8 -j ACCEPT -A OUTPUT -o lo -j ACCEPT 2016-03-31 23:36:14,211 DEBG 'start' stdout output: --------------------