I have implemented a DR solution to use OSSV (v3.0.1) to backup some Linux servers. After the initial baseline the snapvault updates run fine for a couple of days but then the /home partition fails to update with error "Unexpected read select while sending". All other partitions (/var, /boot, /root) continue to update fine. I presume it might be a size related issue since the /home partition gets to 500-600GB in size. This has happened on two servers so far which are both Dell R620 16GB RAM. 6 x 300gb 2.5" sas drives in hardware RAID10.
I have a case open with NetApp who asked for trace logs which I have uploaded (they are massive). We tried setting the volume language to en.UTF-8 in case that was an issue, but to no avail. There is plenty of bandwidth and don't believe its a network issue. The solutions uses a vFiler as a snapvault secondary.
Any help, comments, suggestions would be appreciated as this is really scuppering my new DR strategy.
We have been running OSSV for years. We started having lots of problem when weupgraded to Ontap 8.0.2P6 from 7.3. It works some of the time. We sent NetApp lots of traces from various sites, no resolution to the problem. Not sure if we did something wrong or something change in Ontap 8.0.
Got a response from NetApp saying its bug in processing certain files and the work-a-round is to increase the THROTTLE setting in the wan.cfg file. We were throttling down to 30MB/s.
Alternatively the bug should be fixed in OSSV version 3.0.1P3 so I think we are going to update the OSSV Agent version as this bug affects any OS.
Hope this helps and let me know how you get on.
Just a quick update to say that the updated client hasn't worked. We are continuing to generate logs for engineering to analyze. Interestingly I setup an OSSV backup on a Windows 2003 server and all partitions are successfully backing up for the past few weeks ... is this just a Linux issue I wonder?
Anyone else making progress with this?