Reports 1-1 of 1 Clear search Modify search
CAL (General)
takahiro.yamamoto - 19:15 Monday 16 March 2026 (36588) Print this report
Inspection of a network trouble between CAL and DMG for Low-Latency data transfer
The Low-Latency data transfer has been failing since power was restored following the planned power outage.
So far, I haven't found any issue on L3 level and it seems to be an L2 issue.
To check about L2 settings, we need to access 10GigE Fujitsu switch and ask login info of the company.

-----
The reason for the data transfer failure is that the CAL servers and the DMG servers are unable to communicate with each other. These servers are belongs to DMG LAN and are connected via SR-S352TR1.

At first, we doubted a wrong settings on CAL servers (latest related works are klog#32886 and klog#35374), but two CAL servers were reachable each other and IP settings (ip a) and routing (ip r) seemed to be proper. And also, DMG servers should be fine because the bulk data transfer is working well (of course DMG servers are also reachable each other). I also checked ARP then, MAC address hadn't been resolved either. So CAL servers seemed to locate another place of DMG servers in L2 level. If so, L2 settings are done on SR-S352TR1 because both cal-gst* as CAL servers and hyades-* as DMG servers are directly connected to that network switch.

I don't know the login information to that switch and also DMG folks, so we are now asking the company folks. As what we can do without a login to that switch, I tried to change the connected network ports for cal-gst*. Now cal-gst-1 and -2 are connected to Port#1 and #2. Port#31 and #33-36 are used for UPS and disk storage devices. Port#38-40 are used for iRMC. All of them are connected by UTP cables. Hyades-* and the network switch at Mozumi are connected to Port#49-52 by optical fibers. Even when I used another ports for cal-gst* except Port#32 and #37, a situation didn't changed (CAL servers were reachable each other but unreachable to DMG servers). CAL servers became unreachable each other when I used Port#32 or #37.

For UPS devices and iRMC, dedicated VLAN are assigned. So Port#31-36 and Port#37-40 seems to be assigned another VLAN segment from Port#1-30 and #41-48 at least. I couldn't confirm about Port#49-52 because I didn't have necessary equipments today. I can check them tomorrow by preparing SFP+ module and optical fibers and connecting cal-gst* to Port#49-52.
Comments to this report:
takahiro.yamamoto - 22:33 Tuesday 17 March 2026 (36608) Print this report
Finally, this issue was solved by restoring a vanished VLAN settings on the DMG network switch (klog#36602).
The data transfer for the output of low-latency calibration pipeline was also resumed.
Now, it is nothing more than just a PD dark noise with applying DARM sensing response but the latest LL frame are also available at Kashiwa.
Search Help
×

Warning

×