I've been backing up a Sun StorEdge 5320 for quite some time now and recently they upgraded the firmware from 4.20 M2 to 4.21 M1. Immediately backups began to fail with error 99 in the event viewer. At first, there was a login issue. Sun renamed the "root" user from "adminsitrator" to "admin" which we corrected. That however, did not fix the issue.
I can't see any erros in the NDMP log. The Activity Monitor shows the following:
Supplying updates re actual solutions is the raison d'etre of this forum. You deserve credit for doing what most others don't. Time and time again, suggestions, hints and probable/possible solutions are offered, and the selfish do not post back the actual outcome.
You, on the other hand, didn't even have any solutions/pointers offered, and yet you still came back to post your solution. Thanks again.
I have a non Symantec question. I am looking at replacing my NFS server and am looking at the 5320 as an option. How do you like yours? I have a fairly steady load in a engineering/production environment with just over 3 T of data. Any insite is appreciated. Thank you
The 5320 is nice midrange nas solution at a decent price. It is not intended or marketed as a enterprise solution. It is easily configured and is stable. Firmware upgrades will require a service outage even if you buy the clustered solution.
The device is a bit "quirky" at first, but once you get the hang of it its not too terrible to configure. There's definitely something wrong with their NDMP backup performance. I'm able to get 60MB/sec+ via NFS and barely 20MB/sec backing the device up via NDMP. I'm working with another user from this forum who has these and is having the same issues. Sun support so far has been lack luster. I've got several remote sites currently with these and an apple xraid solution and I'm not sure which way I'd lean in the future. This device is ceratinly feature rich but after buying NBU NDMP Licences the performance has been terrible. I've got one I'm backing up via NFS now because NDMP was taking 3+ days to backup 2TB of data. Hopefully Sun gets us a solution sonner rather than later.
For best performance, you'll want to configure /usr/openv/netbackup/db/config/NET_BUFFER_SZ to 65536. I've found that it will double the performance coming from the 5320 from ~20mb/sec to ~50mb/sec. This is on a Solaris media server - on other hosts or operating systems, YMMV.
Also, if you haven't already done so, tune the NUMBER_DATA_BUFFERS and SIZE_DATA_BUFFERS in the same directory on your media server(s) to their optimum levels as well.
I'd sure like to get that kind of throughput increase. What is your NDMP configuration? In my environment, we have fiber-attached tape drives zoned to be visible to the 5320. I was wondering how tuning the network buffers on the NetBackup server might help with such a configuration.