[Koji, Nikhil, Gautam]
We couldn't get data using python nds2. There seems to have been many problems.
- /frames wasn't mounted on megatron, which was the nds2 server. Solution: added /frames 192.168.113.209(sync,ro,no_root_squash,no_all_squash,no_subtree_check) to /etc/exportfs on fb1, followed by sudo exportfs -ra. Using showmount -e, we confirmed that /frames was being exported.
- Edited /etc/fstab on megatron to be fb1:/frames/ /frames nfs ro,bg,soft 0 0. Tried to run mount -a, but console stalled.
- Used nfsstat -m on megatron. Found out that megatron was trying to mount /frames from old FB (192.168.113.202). Used sudo umount -f /frames to force unmount /frames/ (force was required).
- Re-ran mount -a on megatron.
- Killed nds2 using /etc/init.d/nds2 stop - didn't work, so we manually kill -9'ed it.
- Restarted nds2 server using /etc/init.d/nds2 start.
- Waited for ~10mins before everything started working again. Now usual nds2 data getting methods work.
I have yet to check about getting trend data via nds2, can't find the syntax. EDIT: As Jamie mentioned in his elog, the second trend data is being written but is inaccessible over nds (either with dataviewer, which uses fb as the ndsserver, or with python NDS, which uses megatron as the ndsserver). So as of now, we cannot read any kind of trends directly, although the full data can be downloaded from the past either with dataviewer or python nds2. On the control room workstations, this can also be done with cds.getdata. |