Page 2 of 2 FirstFirst 12
Results 11 to 20 of 20

Thread: ubuntu server disconnection issue

  1. #11
    Join Date
    Mar 2010
    Location
    Squidbilly-Land
    Beans
    Hidden!
    Distro
    Ubuntu

    Re: ubuntu server disconnection issue

    When it doesn't work, run those commands again and see what's different.
    I'd start with the pings ... local, the gateway, then a public IP. If all those work, it isn't a network issue at all.

    Probably just a DNS issue. Search for forum posts about solving that. There are many ways to do that. The way I fixed it forever isn't what most people would do, but you can find those posts in these forums and follow them if you like. Or wait until someone else comes and explains how to make systemd-resolved work.

  2. #12
    Join Date
    May 2024
    Beans
    Hidden!
    Distro
    Ubuntu

    Re: ubuntu server disconnection issue

    Quote Originally Posted by TheFu View Post
    Way to jump to step 45 of troubleshooting this type of issue. There are so many much more likely problems to check first before throwing someone into tcpdump.
    Not really. Packet loss is best diagnosed with a tool to dump packets. Novice users setting static IP assignments and having conflicts on the local network is more common than you think.

  3. #13
    Join Date
    May 2024
    Beans
    Hidden!
    Distro
    Ubuntu

    Re: ubuntu server disconnection issue

    Quote Originally Posted by TheFu View Post
    When it doesn't work, run those commands again and see what's different.
    I'd start with the pings ... local, the gateway, then a public IP. If all those work, it isn't a network issue at all.

    Probably just a DNS issue. Search for forum posts about solving that. There are many ways to do that. The way I fixed it forever isn't what most people would do, but you can find those posts in these forums and follow them if you like. Or wait until someone else comes and explains how to make systemd-resolved work.
    It's not a DNS issue, because as OP said, SSH stops working.

    ping for networking troubleshooting hasn't been useful since about 1998. Most, if not all, secure networks have blocked ICMP traffic years and years ago.

  4. #14
    Join Date
    Nov 2009
    Location
    Catalunya, Spain
    Beans
    14,570
    Distro
    Ubuntu 18.04 Bionic Beaver

    Re: ubuntu server disconnection issue

    I don't want to steal the thread but I might be having the same/similar problem. I am busy these days but I was planning to open a thread of my own soon to see if anyone more expert than me can help. As some might know, I can find my way in ubuntu but this problem has puzzled me for years.

    I think I first noticed it in Ubuntu Server 18.04, then after an in-place upgrade to 20.04 it continued. At that time I was using HDDs in mdadm raid1 for the OS. After that I moved to SSDs in mdadm raid1 for OS and to see if it will help I did a clean new install of 22.04, which I am running now. But the problem is still here.

    What my server does is that it simply stops communicating over the network. All light are still on as normal, fans running, HDD led seems to have ocassional activity, but no network, ping, ssh, nothing. Like you pulled the cable out (but connectivity LED on the NIC and the switch port still shows as normal).

    Obviously without ssh at that moment it is difficult to troubleshoot. I have Supermicro board with IPMI but honestly I don't know too much about IPMI to check if maybe there is local console connection available through it. And I don't have my IPMI port plugged in usually (still working on additional wall RJ-45 socket).

    After a reset it boots normal, and all looks fine. Until the next time. It might "freeze" again the next day, or three weeks later. No apparent rule. CPU temp seems in "normal" range, never seen it above 52-55º C when I can access the server.

    No load of any kind on this server, basically it's for my personal home stuff and ocassional Plex viewing. Really embarassing calling it a server.

    But this network freeze I really can't figure out. As soon as I can I will follow the linked troubleshooting guide to see if I detect something, and I will open my own thread as I said. Just mentioning my experience here, since initially it sounds similar. Network not working suddenly with no apparent reason.
    Darko.
    -----------------------------------------------------------------------
    Ubuntu 18.04 LTS 64bit

  5. #15
    Join Date
    Mar 2010
    Location
    Squidbilly-Land
    Beans
    Hidden!
    Distro
    Ubuntu

    Re: ubuntu server disconnection issue

    Darknod, sometimes integrated NICs fail. I'd drop in a $25 Intel PRO/1000 NIC. BTW, certain types of packets going though certain NICs have been known to brick the NIC. In the 2000s, some Intel Server NICs made in their Mexico plant could be bricked with certain VoIP traffic. Took too many years to figure that out, but I got lucky with a replacement NIC that was made somewhere else and never had the issue again. Then a few years later, I came across an Intel blog article discussing the VoIP traffic/NIC issue. I'd had exactly that model card.

    Over the decades, I've seen very few NIC issues on servers, though the company standard was for redundant connections, always, for everything. When I started running lots of stuff at home, I found that Realtek NICs failed much more often and used more CPU than Intel. Plus the BSDs didn't have good Realtek NIC drivers, so if you use BSD as your way to choose good Linux hardware like I do, then you quickly learn to avoid Realtek and certain other wifi chips (if you use wifi at all).

    I'd much rather have a NIC that fails completely than a NIC that causes data corruption. So having a completely failed NIC is really a good thing, at least in my book. I have a few spare NICs laying around here to help with troubleshooting these sorts of issues. They usually aren't Intel - rather they are the NICs that I replaced with Intel after some issue happened.
    Last edited by TheFu; 2 Weeks Ago at 02:27 AM.

  6. #16
    Join Date
    Feb 2011
    Location
    Coquitlam, B.C. Canada
    Beans
    3,550
    Distro
    Ubuntu Development Release

    Re: ubuntu server disconnection issue

    Quote Originally Posted by eraknix View Post
    For
    Code:
    dmesg |grep eth[0-9]
    :
    Code:
    dmesg: read kernel buffer failed: Operation not permitted
    and with sudo, nothing happens.
    Try using your actual network interface name:

    Code:
    sudo dmesg | grep eno1
    Example, using my NIC name:

    Code:
    doug@s19:~/idle/teo/util2$ sudo dmesg |grep enp3s0
    [    1.293486] igc 0000:03:00.0 enp3s0: renamed from eth0
    [    3.475712] br0: port 1(enp3s0) entered blocking state
    [    3.475717] br0: port 1(enp3s0) entered disabled state
    [    3.475723] igc 0000:03:00.0 enp3s0: entered allmulticast mode
    [    3.475745] igc 0000:03:00.0 enp3s0: entered promiscuous mode
    [    6.377479] igc 0000:03:00.0 enp3s0: NIC Link is Up 1000 Mbps Full Duplex, Flow Control: RX/TX
    [    6.481225] br0: port 1(enp3s0) entered blocking state
    [    6.481229] br0: port 1(enp3s0) entered forwarding state
    Any follow-up information on your issue would be appreciated. Please have the courtesy to report back.

  7. #17
    Join Date
    Feb 2011
    Location
    Coquitlam, B.C. Canada
    Beans
    3,550
    Distro
    Ubuntu Development Release

    Re: ubuntu server disconnection issue

    Quote Originally Posted by currentshaft View Post
    It's not a DNS issue, because as OP said, SSH stops working.

    ping for networking troubleshooting hasn't been useful since about 1998. Most, if not all, secure networks have blocked ICMP traffic years and years ago.
    Was the SSH attempt via IP address or host name?

    With respect, I disagree about "ping". I still think it is one of best low level tools for debugging network issues. But, yes one needs to be aware of potential no responding points. Once I get beyond my gateway, I know that Google's 8.8.8.8 address always responds to ping.

    From another post, I agree with using tcpdump to help.
    Any follow-up information on your issue would be appreciated. Please have the courtesy to report back.

  8. #18
    Join Date
    Jun 2024
    Beans
    6

    Re: ubuntu server disconnection issue

    Hello, I tried for SSH with with local IP address so here it's 192.168.1.87.
    For
    Code:
    sudo dmesg | grep eno1
    :
    The same as before without sudo, I have the same message and with sudo, there's nothing.
    I have to wait the next bug to retest all the commands, I have installed the graphics card this morning, so the next time I will be able to test.
    For
    Code:
    sudo tcpdump -qni any arp
    :
    Code:
    tcpdump: data link type LINUX_SLL2
    tcpdump: verbose output suppressed, use -v[v]... for full protocol decode
    listening on any, link-type LINUX_SLL2 (Linux cooked v2), snapshot length 262144 bytes
    15:07:35.182020 eno1  B   ARP, Request who-has 192.168.1.1 tell 192.168.1.77, length 46
    15:07:40.286167 eno1  B   ARP, Request who-has 192.168.1.1 tell 192.168.1.77, length 46
    15:07:45.391190 eno1  B   ARP, Request who-has 192.168.1.1 tell 192.168.1.77, length 46
    15:07:50.491959 eno1  B   ARP, Request who-has 192.168.1.1 tell 192.168.1.77, length 46
    15:07:51.833359 eno1  In  ARP, Request who-has 192.168.1.87 tell 192.168.1.1, length 46
    15:07:51.833381 eno1  Out ARP, Reply 192.168.1.87 is-at 30:9c:23:ac:99:96, length 28
    15:07:55.597250 eno1  B   ARP, Request who-has 192.168.1.1 tell 192.168.1.77, length 46
    15:08:00.699396 eno1  B   ARP, Request who-has 192.168.1.1 tell 192.168.1.77, length 46
    15:08:05.804398 eno1  B   ARP, Request who-has 192.168.1.1 tell 192.168.1.77, length 46
    15:08:10.908392 eno1  B   ARP, Request who-has 192.168.1.1 tell 192.168.1.77, length 46
    15:08:16.013473 eno1  B   ARP, Request who-has 192.168.1.1 tell 192.168.1.77, length 46
    15:08:16.298147 eno1  In  ARP, Request who-has 192.168.1.87 tell 192.168.1.1, length 46
    15:08:16.298170 eno1  Out ARP, Reply 192.168.1.87 is-at 30:9c:23:ac:99:96, length 28
    15:08:21.117382 eno1  B   ARP, Request who-has 192.168.1.1 tell 192.168.1.77, length 46
    15:08:26.222772 eno1  B   ARP, Request who-has 192.168.1.1 tell 192.168.1.77, length 46
    15:08:31.326779 eno1  B   ARP, Request who-has 192.168.1.1 tell 192.168.1.77, length 46
    15:08:36.431881 eno1  B   ARP, Request who-has 192.168.1.1 tell 192.168.1.77, length 46
    15:08:41.536217 eno1  B   ARP, Request who-has 192.168.1.1 tell 192.168.1.77, length 46
    15:08:46.641679 eno1  B   ARP, Request who-has 192.168.1.1 tell 192.168.1.77, length 46
    15:08:51.747006 eno1  B   ARP, Request who-has 192.168.1.1 tell 192.168.1.77, length 46
    15:08:54.139991 eno1  In  ARP, Request who-has 192.168.1.87 tell 192.168.1.1, length 46
    15:08:54.140016 eno1  Out ARP, Reply 192.168.1.87 is-at 30:9c:23:ac:99:96, length 28
    15:08:56.849744 eno1  B   ARP, Request who-has 192.168.1.1 tell 192.168.1.77, length 46
    ^C
    23 packets captured
    24 packets received by filter
    0 packets dropped by kernel
    192.168.1.77 is my netatmo relay (it's my connected thermostat), I don't know why it's here.
    Tanks for the responses.

  9. #19
    Join Date
    May 2024
    Beans
    Hidden!
    Distro
    Ubuntu

    Re: ubuntu server disconnection issue

    If you run "arp -a" does it show an entry in the cache?

  10. #20
    Join Date
    Nov 2009
    Location
    Catalunya, Spain
    Beans
    14,570
    Distro
    Ubuntu 18.04 Bionic Beaver

    Re: ubuntu server disconnection issue

    FYI in my case dmesg doesn't give any helpful info. Only what I understand to be the initialization of the NIC at OS load:
    Code:
    darko@filesrv:~$ sudo dmesg | grep eno1
    [sudo] password for darko: 
    [    7.872635] igb 0000:03:00.0 eno1: renamed from eth0
    [   14.573649] igb 0000:03:00.0 eno1: igb: eno1 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: RX
    PS. What TheFu suggested about getting a new NIC is good idea although I would like to avoid using one of the few PCIE slots that I have for a NIC. After using simple home boards for my "server" for years I finally got a Supermicro X11-SCL server board with double NIC and it would be shame if it turns out I can't use them.

    I have some USB-to-RJ45 laying around, they might be a way to do some temporary testing acting as new NIC without buying anything yet or populating a PCIE slot.
    Last edited by darkod; 2 Weeks Ago at 09:03 PM.
    Darko.
    -----------------------------------------------------------------------
    Ubuntu 18.04 LTS 64bit

Page 2 of 2 FirstFirst 12

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •