Hello everyone,
I've been getting into doing more and more server setups within my home environment, and that has been slowly building out towards hosting small stuff for friends and family (even email etc). To this end, I picked up a c6100 from a kijiji ad in my area (for those not local, think craigslist)... anyways, it ended up being a DCS node; I followed this guide: How To: Update a Dell C6100 DCS Node (BIOS & BCM) | Server Network Tech and updated the bios and BCM to the version posted in that listing (BIOS version 1.71 and 1.33 BCM).
3 of 4 nodes updated fine and have no issues; I've been able to add 2 of the 3 working nodes to an ESXi cluster and it's working great, the 3rd I left out for testing alongside the failed node.
By failed, I don't mean failed. This issue existed before and after the update; update went fine, I'll explain.
-- History (before powering on the nodes) --
When I first got the system I inspected it by removing each node and examining it for visible damage; I didn't find anything but on node 1 there was an odd memory configuration. I noted that the other nodes had all their memory in the white slots; but one dimm on one node was in a black slot. I noticed and moved it from the black slot to the appropriate white slot. I thought nothing more of it at the time.
-- The update and the issue --
I updated 2 nodes easily (nodes 3 and 4), the next node I tried was node 1; this happened to be the node I noted the inconsistent memory configuration with earlier. The node posted normally, BCM was accessible, everything appeared to work, but when it came to booting up the node, just like every other node had booted (device selection and ultimately OS start) the node came to a blinking cursor, and refused to proceed. No error, nothing, just a flashing cursor.
-- The update completed --
I completed the update by removing all the RAM for one CPU, and one stick for the second. (the front CPU had 2x4GB of memory) - the node booted a bit slowly, but it seemed to run O.K. for the most part, and I was able to flash the BIOS and BCM without any further trouble. BIOS is now 1.71 and BCM now 1.33. I moved on, and shelved the issue for the day. Afterall, I still had Node 2 to flash, so I did that.
-- The issue continues --
I re-tested the system with all different memory configurations, I even pulled memory from another node to test in it, the results were consistent: With 6x4GB RAM, the system would stop at the flashing cursor, with at least 1 DIMM missing from each CPU, it would boot fine, but slowly.
Right now I've kind of had it with this sled, I'm currently looking into buying a new sled, or mainboard or something to replace it with. I'm here as a last-attempt to brainstorm some good ideas before I replace the motherboard.
Has anyone seen behavior like this from their c6100 before?
Specs:
C6100, 4x node - identical config per node:
2xL5520
6x4GB Memory
Model: XS23-TY3
I've been getting into doing more and more server setups within my home environment, and that has been slowly building out towards hosting small stuff for friends and family (even email etc). To this end, I picked up a c6100 from a kijiji ad in my area (for those not local, think craigslist)... anyways, it ended up being a DCS node; I followed this guide: How To: Update a Dell C6100 DCS Node (BIOS & BCM) | Server Network Tech and updated the bios and BCM to the version posted in that listing (BIOS version 1.71 and 1.33 BCM).
3 of 4 nodes updated fine and have no issues; I've been able to add 2 of the 3 working nodes to an ESXi cluster and it's working great, the 3rd I left out for testing alongside the failed node.
By failed, I don't mean failed. This issue existed before and after the update; update went fine, I'll explain.
-- History (before powering on the nodes) --
When I first got the system I inspected it by removing each node and examining it for visible damage; I didn't find anything but on node 1 there was an odd memory configuration. I noted that the other nodes had all their memory in the white slots; but one dimm on one node was in a black slot. I noticed and moved it from the black slot to the appropriate white slot. I thought nothing more of it at the time.
-- The update and the issue --
I updated 2 nodes easily (nodes 3 and 4), the next node I tried was node 1; this happened to be the node I noted the inconsistent memory configuration with earlier. The node posted normally, BCM was accessible, everything appeared to work, but when it came to booting up the node, just like every other node had booted (device selection and ultimately OS start) the node came to a blinking cursor, and refused to proceed. No error, nothing, just a flashing cursor.
-- The update completed --
I completed the update by removing all the RAM for one CPU, and one stick for the second. (the front CPU had 2x4GB of memory) - the node booted a bit slowly, but it seemed to run O.K. for the most part, and I was able to flash the BIOS and BCM without any further trouble. BIOS is now 1.71 and BCM now 1.33. I moved on, and shelved the issue for the day. Afterall, I still had Node 2 to flash, so I did that.
-- The issue continues --
I re-tested the system with all different memory configurations, I even pulled memory from another node to test in it, the results were consistent: With 6x4GB RAM, the system would stop at the flashing cursor, with at least 1 DIMM missing from each CPU, it would boot fine, but slowly.
Right now I've kind of had it with this sled, I'm currently looking into buying a new sled, or mainboard or something to replace it with. I'm here as a last-attempt to brainstorm some good ideas before I replace the motherboard.
Has anyone seen behavior like this from their c6100 before?
Specs:
C6100, 4x node - identical config per node:
2xL5520
6x4GB Memory
Model: XS23-TY3
