This post is more than 5 years old
10 Posts
1
2489
XIO XMS port speed requirement
Hi All, have a question that I'm hoping the community can help answer.
Customer networking team configured the XMS and Storage Controllers at 100Mb Full duplex and the health checks (pre-upgrade) have flagged this.
Support have provided the article that discusses this with recommendations to change to 1000Mb. Also that the site install doc states all interfaces at 1000Mb. Customer network switches within the define zone are not capable of this, however there may be an option to move only the Storage Controller interfaces into the Prod Zone and firewalled.
Question I have is if the health check script only checks for the SCs at 1000Mb or also the XMS? It seems to initially fail at the first SC. From your experiences, is there a requirement for the XMS to also have that speed?
Thanks for your time,
Daniel
Anonymous
5 Practitioner
5 Practitioner
•
274.2K Posts
0
June 26th, 2017 13:00
Trouble is that there is no unofficial option/answer to this question. The problem is in the transfer of large files, include the upgrade image during NDU and the creation of regular debug logs which are then uploaded to EMC via SYR or ConnectEMC. By limiting the link rate to 100Mb, timeouts may occur which impact NDU and log creation/collection. Unofficially, as long as the link is healthy (no errors), all other mgmt. functions should not be impacted by the lower link rate but why take the risk?
scotthoward
64 Posts
2
June 7th, 2017 21:00
Officially the XMS and the storage controllers must all be connected using GigE, and the link(s) between the XMS and the SC's must also be (at least) GigE.
Anything slower, and the upgrade can fail due to the time it takes to copy the upgrade file to each of the storage controllers.
D_Rep
10 Posts
0
June 8th, 2017 14:00
Hi Scott, thanks for your response. I suppose I was looking for an unofficial option, in terms of what is actually possible with the XMS being connected with 100Mb and if there was it would pass the health check and ultimate upgrade.
D_Rep
10 Posts
0
June 26th, 2017 18:00
Thanks for the further reply in how in how it works and dependencies on the connections.Agree I don't want to take the risk but this customer install is on switches that are not 1Gb capable and to be so, they have to go through an entire switch refresh and these are not in scope short term. To be so means additional cost and about a 4 month lead time and in the meantime the arrays have to undergo upgrade to remain at target code. Hence looking at options around what is possible, especially where the health check script errors on the SC port speed initially.