> Unfortuanetly, your description is not very clear. > Can you be a bit more precise? > > Your actual problem seems to be, that you cannot > import your old zpool on a new version of > (Open)Solaris, right? This is actually more of a ZFS > problem, so I'd suggest to direct your question to > the zfs discussion on this forum - you will be > getting much better help there. > > Anyway… when I suffered from a zpool that wouldn't > import and that crashed my server whenever I tried to > import it, I have been given a couple of advices: > > a) have the system generate a core dump for analysis > b) try to mount the zpool read-only, since it might > be one ore more file systems on the pool, that causes > these crashes > > Cheers, > budy I have learned to live with the can't import zfs. Last night I did a little more testing with my problem of running out of memory and hitting swap without anything running except a nfs file transfer.
So last night I did a zfs send to a local disk of 180Gbyte zfs filesystem. This went well without a issue. Next I did a dd of that 180Gbyte file from my local zfs filesystem to a nfs mounted zfs filesystem. After about 80Gbyte transfered the system started swapping out "w column in vmstat has over 200 apps swapped out all 4Gbytes of memory gone and system performance went into the toilet. Nothing else except the standard Solaris OS stuff running. No reason for it to consume so much memory. Something in either the network driver or zfs over nfs is sucking all my memory up. -- This message posted from opensolaris.org _______________________________________________ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
