Learn about Centmin Mod LEMP Stack today
Become a Member

Vultr SSH time outs with Vultr

Discussion in 'Virtual Private Server (VPS) hosting' started by Josephm, Sep 13, 2014.

  1. Josephm

    Josephm Active Member

    126
    43
    28
    Aug 26, 2014
    Ratings:
    +45
    Local Time:
    2:53 PM
    1.9.5
    10.0.21
    @eva2000, do you have any network issue with Vultr? After made fresh installation OS & centmindmod, I restarted my vpses, and all is running well. However; after a day, or sometimes some hours later, my servers have time out error "The connection has timed out", I tried proxy or TOR but had same result. (console from Vultr is accessible, but web server to my vps is not). I had to restart vps again to see my webserver running.
    Could you try then keep that vps running for a day?
    I tested 2 vpses in Vultr in Illinois and New Jersey, they had same result. I can't find the contact support to ask...
    My test vps at Ramnode is quite good, not have same connection error...
     
    • Like Like x 1
  2. eva2000

    eva2000 Administrator Staff Member

    30,196
    6,789
    113
    May 24, 2014
    Brisbane, Australia
    Ratings:
    +10,144
    Local Time:
    5:53 PM
    Nginx 1.13.x
    MariaDB 5.5
    • Like Like x 1
  3. Josephm

    Josephm Active Member

    126
    43
    28
    Aug 26, 2014
    Ratings:
    +45
    Local Time:
    2:53 PM
    1.9.5
    10.0.21
    Thanks for your info, after the benchmark, I wanted Vultr but now have to come back Ramnode anyway :(
     
  4. eva2000

    eva2000 Administrator Staff Member

    30,196
    6,789
    113
    May 24, 2014
    Brisbane, Australia
    Ratings:
    +10,144
    Local Time:
    5:53 PM
    Nginx 1.13.x
    MariaDB 5.5
    tried a different Vultr region ?
     
  5. Josephm

    Josephm Active Member

    126
    43
    28
    Aug 26, 2014
    Ratings:
    +45
    Local Time:
    2:53 PM
    1.9.5
    10.0.21
    Yes, I already tried to both new jersey & Illinois and had same timeout issues. I really do not want my web to be unstable...
     
  6. eva2000

    eva2000 Administrator Staff Member

    30,196
    6,789
    113
    May 24, 2014
    Brisbane, Australia
    Ratings:
    +10,144
    Local Time:
    5:53 PM
    Nginx 1.13.x
    MariaDB 5.5
    Strange might want to contact Vultr support via your my.vultr.com area
     
  7. Josephm

    Josephm Active Member

    126
    43
    28
    Aug 26, 2014
    Ratings:
    +45
    Local Time:
    2:53 PM
    1.9.5
    10.0.21
    Just tested Vultr in Illinois with Centos 6.5_64 and seems running fine with network for hours
    I opened ticket to Vultr and they answered problem has been fixed, I'll try again with Centos 7 with your beta script if network is working fine.
     
    • Like Like x 2
  8. Josephm

    Josephm Active Member

    126
    43
    28
    Aug 26, 2014
    Ratings:
    +45
    Local Time:
    2:53 PM
    1.9.5
    10.0.21
    It seems that time-out problem comes from Centos 7 & centminmod 8beta...
    I tested with Centos 6 & centminmod 7, it worked fine.
    Cenminmod beta should be tested in more other cloud vps before to be stable....
     
  9. eva2000

    eva2000 Administrator Staff Member

    30,196
    6,789
    113
    May 24, 2014
    Brisbane, Australia
    Ratings:
    +10,144
    Local Time:
    5:53 PM
    Nginx 1.13.x
    MariaDB 5.5
    strange indeed, CentOS 7 + centmin mod .08 beta works fine for RamNode and DigitalOcean. Will test Vultr too
     
  10. Josephm

    Josephm Active Member

    126
    43
    28
    Aug 26, 2014
    Ratings:
    +45
    Local Time:
    2:53 PM
    1.9.5
    10.0.21
    Just found that something has removed my route table in Vultr vps; I installed everything Centos 7 & centminmod beta fresh
    Before:
    [​IMG]
    After:
    [​IMG]
    I did not have this issue in centos 6.5 & centminmod 7 stable.
    Do you have any ideas why it's removed and how to see the log of it?
    Temporary solution for me now to restart vps then the route automatically added like Before. After some minutes, it happens again.
    Or I used "ifdown eth0" and "ifup eth0" from Vultr console window, it's working again for a while...
     
    Last edited: Sep 18, 2014
    • Like Like x 1
  11. Josephm

    Josephm Active Member

    126
    43
    28
    Aug 26, 2014
    Ratings:
    +45
    Local Time:
    2:53 PM
    1.9.5
    10.0.21
    Here is my eth0 config:
    Code:
    [root@vu7 network-scripts]# cat /etc/sysconfig/network-scripts/ifcfg-eth0
    HWADDR="52:54:00:9A:22:9E"
    TYPE="Ethernet"
    BOOTPROTO="dhcp"
    DEFROUTE="yes"
    PEERDNS="yes"
    PEERROUTES="yes"
    IPV4_FAILURE_FATAL="no"
    IPV6INIT="yes"
    IPV6_AUTOCONF="yes"
    IPV6_DEFROUTE="yes"
    IPV6_PEERDNS="yes"
    IPV6_PEERROUTES="yes"
    IPV6_FAILURE_FATAL="no"
    NAME="eth0"
    UUID="7cd72280-8d4a-44ef-a73a-e3499ded7efc"
    ONBOOT="yes"
    
    
    Does dhcpclient impact to this interface?
    Vultr status:
    Code:
    [root@vu7 network-scripts]# systemctl status NetworkManager
    NetworkManager.service - Network Manager
      Loaded: loaded (/usr/lib/systemd/system/NetworkManager.service; enabled)
      Active: active (running) since Thu 2014-09-18 07:05:34 UTC; 2h 17min ago
     Main PID: 593 (NetworkManager)
      CGroup: /system.slice/NetworkManager.service
      ├─ 593 /usr/sbin/NetworkManager --no-daemon
      └─3121 /sbin/dhclient -d -sf /usr/libexec/nm-dhcp-helper -pf /var/run/dhclient-eth0.pid -lf /var/lib/NetworkManage...
    Sep 18 09:22:17 vu7 dhclient[3121]: dhclient.c:2676: Failed to send 300 byte long packet over fallback interface.
    Sep 18 09:22:17 vu7 NetworkManager[593]: DHCPREQUEST on eth0 to 107.191.50.16 port 67 (xid=0x1a3829aa)
    Sep 18 09:22:17 vu7 NetworkManager[593]: send_packet: Operation not permitted
    Sep 18 09:22:17 vu7 NetworkManager[593]: dhclient.c:2676: Failed to send 300 byte long packet over fallback interface.
    Sep 18 09:22:30 vu7 dhclient[3121]: DHCPREQUEST on eth0 to 107.191.50.16 port 67 (xid=0x1a3829aa)
    Sep 18 09:22:30 vu7 dhclient[3121]: send_packet: Operation not permitted
    Sep 18 09:22:30 vu7 dhclient[3121]: dhclient.c:2676: Failed to send 300 byte long packet over fallback interface.
    Sep 18 09:22:30 vu7 NetworkManager[593]: DHCPREQUEST on eth0 to 107.191.50.16 port 67 (xid=0x1a3829aa)
    Sep 18 09:22:30 vu7 NetworkManager[593]: send_packet: Operation not permitted
    Sep 18 09:22:30 vu7 NetworkManager[593]: dhclient.c:2676: Failed to send 300 byte long packet over fallback interface.
    
    But my RAMnode is fine, in the same way I installed:
    Code:
    [root@ramnode ~]# systemctl status NetworkManager
    NetworkManager.service - Network Manager
      Loaded: loaded (/usr/lib/systemd/system/NetworkManager.service; enabled)
      Active: active (running) since Sat 2014-09-13 07:42:55 UTC; 5 days ago
     Main PID: 576 (NetworkManager)
      CGroup: /system.slice/NetworkManager.service
      ├─576 /usr/sbin/NetworkManager --no-daemon
      └─694 /sbin/dhclient -d -sf /usr/libexec/nm-dhcp-helper -pf /var/r...
    Warning: Journal has been rotated since unit was started. Log output is incomplete or unavailable.
    
    Any help from you is really appreciate.
     
    Last edited: Sep 18, 2014
    • Like Like x 1
  12. Josephm

    Josephm Active Member

    126
    43
    28
    Aug 26, 2014
    Ratings:
    +45
    Local Time:
    2:53 PM
    1.9.5
    10.0.21
    Found interesting messages log about avahi-daemon:
    [​IMG]

    And avahi-daemon status:
    [​IMG]

    Should I stopped or removed it from system? As in RAMnode this service is running without issue.
     
    Last edited: Sep 18, 2014
    • Like Like x 1
  13. eva2000

    eva2000 Administrator Staff Member

    30,196
    6,789
    113
    May 24, 2014
    Brisbane, Australia
    Ratings:
    +10,144
    Local Time:
    5:53 PM
    Nginx 1.13.x
    MariaDB 5.5
    Might need to whitelist ports for dhclient for UDP ports 67 and 68 in CSF Firewall

    step 1 - add port 67 and 68 to UDP_IN/UDP_OUT to /etc/csf/csf.conf

    Code:
    grep UDP_ /etc/csf/csf.conf
    
    UDP_IN = "111,2049,1110,20,21,53"
    UDP_OUT = "111,2049,1110,33434:33523,20,21,53,113,123"
    so it becomes
    Code:
    UDP_IN = "111,2049,1110,20,21,53,67,68"
    UDP_OUT = "111,2049,1110,33434:33523,20,21,53,113,123,67,68"
    step 2 - remove ports 67 and 68 from DROP_NOLOG in /etc/csf/csf.conf

    Code:
    DROP_NOLOG = "67,68,111,113,135:139,445,500,513,520"
    so it becomes
    Code:
    DROP_NOLOG = "111,113,135:139,445,500,513,520"
    step 3 - restart CSF Firewall

    Code:
    csf -r
     
    • Like Like x 2
  14. eva2000

    eva2000 Administrator Staff Member

    30,196
    6,789
    113
    May 24, 2014
    Brisbane, Australia
    Ratings:
    +10,144
    Local Time:
    5:53 PM
    Nginx 1.13.x
    MariaDB 5.5
  15. Josephm

    Josephm Active Member

    126
    43
    28
    Aug 26, 2014
    Ratings:
    +45
    Local Time:
    2:53 PM
    1.9.5
    10.0.21
    Just follow your steps, hope it will be fine ;) Thank you so much :p
     
    • Like Like x 1
  16. eva2000

    eva2000 Administrator Staff Member

    30,196
    6,789
    113
    May 24, 2014
    Brisbane, Australia
    Ratings:
    +10,144
    Local Time:
    5:53 PM
    Nginx 1.13.x
    MariaDB 5.5
    Yes please let us know how it does and much appreciate the feedback and troubleshooting info (y)
     
  17. RoldanLT

    RoldanLT Well-Known Member

    3,901
    949
    113
    May 25, 2014
    Phillipines
    Ratings:
    +1,298
    Local Time:
    3:53 PM
    1.11
    10.2
    So this is specific only for CentOS 7?
     
  18. Josephm

    Josephm Active Member

    126
    43
    28
    Aug 26, 2014
    Ratings:
    +45
    Local Time:
    2:53 PM
    1.9.5
    10.0.21
    I already changed it and will give feedbacks in hours later to test;
    Come back to my RamNode VPS; CSF config is same, but it does not have this problem:
    Code:
    [root@ramnode ~]# grep UDP_ /etc/csf/csf.conf
    UDP_IN = "111,2049,1110,20,21,53"
    UDP_OUT = "111,2049,1110,33434:33523,20,21,53,113,123"
    
    [root@ramnode ~]# grep DROP_NOLOG /etc/csf/csf.conf
    DROP_NOLOG = "67,68,111,113,135:139,445,500,513,520"
    
    Not sure why, I asked Vultr but he had no ideas at all.
     
    • Like Like x 1
  19. Josephm

    Josephm Active Member

    126
    43
    28
    Aug 26, 2014
    Ratings:
    +45
    Local Time:
    2:53 PM
    1.9.5
    10.0.21
    My Vultr VPS is still running, seems that it's ok now ;)
     
    • Like Like x 1
  20. eva2000

    eva2000 Administrator Staff Member

    30,196
    6,789
    113
    May 24, 2014
    Brisbane, Australia
    Ratings:
    +10,144
    Local Time:
    5:53 PM
    Nginx 1.13.x
    MariaDB 5.5
    awesome :D