topic MSA 2050 replication issues in MSA Storage

MSA 2050 replication issues

agw2 — Tue, 06 Jun 2023 06:56:33 GMT

Hello-

I have two large MSAs that replicate data, one way. The replication runs and completes sucessfully. The problem is that when replication starts, the video server that uses the MSA's storage, starts throwing errors like "low resources" and "database lost" errors. This usually lasts for about 5 minutes into the replication and then all is fine until the replication starts again on it's next scheduled cycle. On the video server, I have done pretty much everything to rule this out to being the problem. I have replaced the hardware, reinstalled the stoftware including the database and OS and still have a problem when replication starts on the MSA. The video server hardware is running 128 gigs of ram and 56 CPU cores.

Can anyone shed some light as to what is going on here?

Query: MSA 2050 replication issues

support_s — Wed, 31 May 2023 22:20:51 GMT

System recommended content:

1. HPE MSA 2050 Storage - Troubleshooting

2. HPE MSA 2050 Storage - Supported Configurations

Please click on "Thumbs Up/Kudo" icon to give a "Kudo".

Thank you for being a HPE valuable community member.

Re: MSA 2050 replication issues

JonPaul — Fri, 02 Jun 2023 17:33:19 GMT

@agw2
Video storage tends to have a very low threshold for latency. Is the replication task causing a performance drop that is interfering with the server standard tasks?
During a replication there are tasks to complete
- taking snapshots
- copying data
While a snapshot is usually very quick in virtual storage it does start a Redirect of the standard IO.
Here is a whitepaper on replication process for MSA 2050: https://www.hpe.com/psnow/doc/4aa1-0977enw
You also could have a port bottleneck, is replication and I/O using the same MSA Host Port?

Re: MSA 2050 replication issues

agw2 — Thu, 08 Jun 2023 11:30:28 GMT

@JonPaul In Event Viewer, I've noticed on the Video server connected to the MSA, when replication occurs the error 147 "An IO took more than 30000 ms to complete" and also error 149 "In the past 3600 seconds we had IO failures". There may be a bug in the MSA.