Repeated LAN Link Down / Link Up messages in syslog

Questions pertaining to Power settings, Auto Block, Permissions, User Quotas and Email alerts may be posted here.
Forum rules
Synology Community is the new platform for the enthusiasts' interaction, and it will soon be available to replace the Forum.
tinue
Trainee
Trainee
Posts: 16
Joined: Fri Oct 30, 2015 3:02 pm

Re: Repeated LAN Link Down / Link Up messages in syslog

Unread post by tinue » Mon Sep 03, 2018 1:01 pm

My system shows the same error. In addition, the CPU is permanently busy. When I ssh onto the system, and show the log ('dmesg'), I see this:

Code: Select all

[177494.347147] IPv6: ADDRCONF(NETDEV_UP): eth3: link is not ready
[177496.122284] igb 0000:00:14.3 eth3: igb: eth3 NIC Link is Up 10 Mbps Full Duplex, Flow Control: RX/TX
[177496.132669] IPv6: ADDRCONF(NETDEV_CHANGE): eth3: link becomes ready
[177497.666543] igb 0000:00:14.3: changing MTU from 1500 to 9000
[177499.591070] igb 0000:00:14.3 eth3: igb: eth3 NIC Link is Up 10 Mbps Full Duplex, Flow Control: RX/TX
[177510.751047] igb 0000:00:14.3: changing MTU from 9000 to 1500
[177511.394738] IPv6: ADDRCONF(NETDEV_UP): eth3: link is not ready
[177513.110858] igb 0000:00:14.3 eth3: igb: eth3 NIC Link is Up 10 Mbps Full Duplex, Flow Control: RX/TX
[177513.121251] IPv6: ADDRCONF(NETDEV_CHANGE): eth3: link becomes ready
[177514.668171] igb 0000:00:14.3: changing MTU from 1500 to 9000
[177516.578613] igb 0000:00:14.3 eth3: igb: eth3 NIC Link is Up 10 Mbps Full Duplex, Flow Control: RX/TX
[177527.798222] igb 0000:00:14.3: changing MTU from 9000 to 1500
This happens only on one link. This link is directly connected with a patch cable to my old Synology. I use this for backup purposes. As soon as I turn on the old Synology, the problem disappears:

Code: Select all

[177687.564242] igb 0000:00:14.3 eth3: igb: eth3 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: RX/TX
So in my case the solution is simple: When I turn off the old Synology, I have to physically unplug the Ethernet cable from LAN 3.

Synology support declared my machine to be broken and recommended to get it replaced. I wanted to make another backup before I send in the machine, which is when I discovered the above pattern.

Squozen
Guru
Guru
Posts: 1548
Joined: Wed Jan 09, 2013 1:35 am

Re: Repeated LAN Link Down / Link Up messages in syslog

Unread post by Squozen » Mon Sep 03, 2018 4:59 pm

Does that problem go away if you have a switch between the old Synology and the new one?

tinue
Trainee
Trainee
Posts: 16
Joined: Fri Oct 30, 2015 3:02 pm

Re: Repeated LAN Link Down / Link Up messages in syslog

Unread post by tinue » Tue Sep 04, 2018 9:13 am

Squozen wrote:
Mon Sep 03, 2018 4:59 pm
Does that problem go away if you have a switch between the old Synology and the new one?
Yes, it does go away.
I wanted the fastest possible speed for the backup, and I wanted to use jumbo packets. That's why I connect the two Synologies directly. My switch is rather cheap and not very intelligent, unfortunately.

Squozen
Guru
Guru
Posts: 1548
Joined: Wed Jan 09, 2013 1:35 am

Re: Repeated LAN Link Down / Link Up messages in syslog

Unread post by Squozen » Tue Sep 04, 2018 9:28 am

Do you actually see much improvement from jumbo frames? Enough to make it worth the hassle?

tinue
Trainee
Trainee
Posts: 16
Joined: Fri Oct 30, 2015 3:02 pm

Re: Repeated LAN Link Down / Link Up messages in syslog

Unread post by tinue » Tue Sep 04, 2018 12:25 pm

Now that you mention it: No, not at all. The setup remained after an intense troubleshooting session related to backup.

When I setup Hyper Backup / Hyper Backup Vault, it was extremely slow and utilized maybe 1/3 or less of the theoretical bandwidth of the network. I even had a support ticket open to get this analyzed, without getting an improvement.

I used the backup when I switched to BTRFS on my main Synology. I remember that the restore of about 5TB took days to complete on a Gigabit Ethernet connection. One of many things that I tried was to use a dedicated link, with Jumbo Frames enabled. It may have helped a tiny bit, but the slow speed was mainly due to the Hyper Backup software, not because of the network.

Over time, Hyper Backup has become significantly faster. I never repeated any measurements so I can't say how quick it is with the current release.

Thanks for the question, it prompted me to rethink the setup, and I'll remove the cable now ;)

AMT
I'm New!
I'm New!
Posts: 1
Joined: Mon Sep 10, 2018 12:21 am

Re: Repeated LAN Link Down / Link Up messages in syslog

Unread post by AMT » Mon Sep 10, 2018 12:25 am

turncloud wrote:
Tue Aug 28, 2018 4:08 pm
We have just hit this problem again (Link Up / Link Down) on the replacement NAS whilst the customer's original one is with Synology for repair (we hold a hot spare). Exactly the same symptoms.

We have just disabled DHCP and moved the DHCP function to the router and the unit is starting to behave again. Will watch it now but CPU use has dropped from 99% to 34% and no more link up / link down messages yet.
Do you have a managed switch that has logs you can check? I have seen link up/down issues with Spanning Tree configurations that are bad.

brighp
I'm New!
I'm New!
Posts: 1
Joined: Sun Feb 05, 2017 1:20 pm

Re: Repeated LAN Link Down / Link Up messages in syslog

Unread post by brighp » Sun Sep 16, 2018 7:16 pm

Just noticed hundreds of LAN 1/LAN 2 Disconnected/Connected messages. Only just noticed because the server has been performing perfectly. Bonded connection. System has been up for nearly a year with only reboots following every DSM update. No changes made to cables, switches, routers or infra of any sort during the past year.

On checking the log (which I never clear down) the problem first started happening on 26th June 2018 following a DSM update. Unfortunately I can't say which particular update other than I update within a week or so of receiving notification of an update. Currently on version DSM 6.2-23739 Update 2.

The problem is worsening in frequency. For the month to date LANs 1 & 2 are up and down repeatedly with gaps of as little as 5mins but never more than 30mins apart. As per the log below (which shows the first time the problem occurred on the DS716) the links are down for only 2 or 3 seconds before they're up again.

Info System 29/06/18 22:07:10 SYSTEM [LAN 1 (slave)] link up.
Info System 29/06/18 22:07:08 SYSTEM [LAN 1 (slave)] link down.
Info System 29/06/18 22:07:08 SYSTEM [LAN 1 (slave)] link down.
Info System 29/06/18 22:07:08 SYSTEM [LAN 2 (slave)] link up.
Info System 29/06/18 11:27:46 SYSTEM System successfully registered [xxx.xxx.xxx.xxx] to [xxxxxx] in DDNS server [xxxxx].
Info System 29/06/18 00:00:57 SYSTEM [LAN 2 (slave)] link up.
Info System 29/06/18 00:00:55 SYSTEM [LAN 2 (slave)] link down.
Info System 29/06/18 00:00:55 SYSTEM [LAN 2 (slave)] link down.
Info System 29/06/18 00:00:55 SYSTEM [LAN 1 (slave)] link up.
Info System 28/06/18 11:28:02 SYSTEM System successfully registered [xxx.xxx.xxx.xxx] to [xxxxxx] in DDNS server [xxxxx].
Info System 28/06/18 10:56:47 SYSTEM [LAN 1 (slave)] link up.
Info System 28/06/18 10:56:45 SYSTEM [LAN 1 (slave)] link down.
Info System 28/06/18 10:56:45 SYSTEM [LAN 1 (slave)] link down.
Info System 28/06/18 10:56:45 SYSTEM [LAN 2 (slave)] link up.
Info System 27/06/18 17:57:23 SYSTEM Package [Universal Search] has been successfully updated.
Info System 27/06/18 11:28:04 SYSTEM System successfully registered [xxx.xxx.xxx.xxx] to [xxxxxx] in DDNS server [xxxxx].
Info System 26/06/18 11:40:22 SYSTEM System successfully registered [xxx.xxx.xxx.xxx] to [xxxxxx] in DDNS server [xxxxx].
Info System 26/06/18 11:34:29 SYSTEM [Cloud Station] service was started.
Info System 26/06/18 11:33:40 SYSTEM [USB Copy] service was started.
Info System 26/06/18 11:33:15 SYSTEM [Cloud Sync] service was started.
Info System 26/06/18 11:25:25 SYSTEM [Cloud Sync] service was stopped.
Info System 26/06/18 11:25:17 SYSTEM [Cloud Station] service was stopped.
Info System 26/06/18 11:25:16 SYSTEM [USB Copy] service was stopped.
Info System 26/06/18 11:25:07 System started counting down to reboot.
Info System 26/06/18 11:24:56 Update was complete.

Finally, I also have a DS414Slim, which I generally use for testing/extended backup. Same DSM version. That too is reporting the same issue that started occurring around the same date as the DS716. However, not at the same frequency as that only auto boots once a day for an hour just to keep things alive. In all other respects the problem is similar to the DS716.

I reckon it's a DSM change; a heartbeat/keep-alive has been made more sensitive. It would be good to get a definitive answer from Synology as to whether they're seeing the same issue and what the fix time is. For me, at present it's an irritant rather than a critical problem.

Systems: DS716+II with DX513 extension & DS414Slim

Post Reply

Return to “System Management”