CHATS: base Messages: 5 Entries..

18may07

With Ian's Formaldehyde coming Monday with it's supposed 7A 1-phase load,
plus his desire for airconditioning, I decided to recheck the base power.


With A/C, battery charger, lights, Motosat and several PC's running:
Red-Leg: 16A
Blk-Leg: 18A
50A breaker in building, so we should be OK for Ian's

15may07....18may07

We've been having troubles with the pocketecs especially at hftower,
but also at west overnight.   On a regular basis we need to make
sure a file-system-check is done on these devils.

Gordon's 'copy' script (do this first)
-----------------------
   copy_usbdisk.sh

Gordon's 'checking' script: (do this second)
-----------------------
   bb_fsck.sh
	This will take several hours to do, and you should first
	clean off the media using the script above.
	It essentially does a
	fsck -c -c with both a read plus a 'read/write' diagnostic
	of bad blocks.   This should help overcome the difficulties
	with these devices if regularly performed.
	Gordon's script also creates a log of activity (see below)
	and if you 'ps -elf | grep fsck' during the operation, you'll
	get something like this.

0 S aster    15607 25915  0  79   0 -  1102 wait   09:37 pts/8    00:00:00 /bin/sh /usr/local/isff/projects/CHATS/ISFF/scripts/bb_fsck.sh
4 S root     15644 15607  0  81   0 -  1039 -      09:37 pts/8    00:00:00 fsck -c -c -p /dev/sdd1
0 S aster    15645 15607  0  79   0 -   909 pipe_w 09:37 pts/8    00:00:00 tee -a /usr/local/isff/projects/CHATS/ISFF/logs/usbdisk_ptec11_20070518_093729_fsck.log
4 S root     15646 15644  0  78   0 -  1488 -      09:37 pts/8    00:00:00 fsck.ext3 -c -c -p /dev/sdd1


To check the disk by hand:
--------------------------
1) connect on usb port, while watching tail -f /var/log/messages
   for the device to be recognized.
   NOTE: make certain the usb cable is secure, these can be loose!
2) Konquerer will come up: cancel that
3) mount
   to see whether it is mounted or not.
   should come up as /media/usbdisk
   umount if it is.
4) su root
5) run the file system check on appropriate device: fsck.ext3 /dev/sdc1
   report should look like this....
e2fsck 1.38 (30-Jun-2005)
usbdisk: recovering journal
usbdisk has been mounted 119 times without being checked, check forced.
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information
usbdisk: 16/7325696 files (18.8% non-contiguous), 280011/14651272 blocks

Inventory: 8 pocketecs, 2 spares here
-------------------------------------
ptec#2	15may07:OK reinstalled at west
ptec#11	15may07:OK reinstalled at hftower but subsequent errors
	18may07:...
ptec#6  15may07:OK but subsequent errors until 'bb_fsck'
	17may07:OK
ptec#7	The 'good unit' used for many days in array
	and put into hftower on 5/17
ptec#1	

13may07 18:00pdt

Stopped dsm, etc.
Moved power strip with 'aster latitude, isff2, AP24, D-link'
from raw power plug to UPS plug.
Last night we saw data outage at roughly 2-3am local due
to a power hit.   The AP24, etc died and r.t. data was
halted.   The aster latitude stayed up due to internal
batt. but the dlink, etc was out.   The adams appeared
to stay up based on file create dates but may have had
some loss based on raw file sizes....not yet fully
investigated.
today had 'poor' wind dir. so we decided to do this now.

It will take 40 days to fill up east's pocketec, which has the
biggest data files I believe (1.5 Gbyte/day).  pocketecs are 60 Gbyte.

The network is running well. I think we have 100% recovery on
the base system.

So, copying pocketecs is relatively low priority.  Maybe once a week on
the tower, and on config changes on the array.

I need to create the script to copy the data and clear the pocketec.

The hotfilms are another matter, since we're not sending that data
over the net. The 3D system generates 6.2 Gbytes/day if left
running continuously.  So one should swap that drive every week
or so if we run it often.  The 5 1D probes generate 4.8 Gbyte/day.

There should have been two spare pocketecs in the hotfilm shipment.

If you think the trailer is safe from theft, then perhaps you
could leave isff7 connected, and we can do a daily copy from
isff1 to 7.

In order to recalculate covars:

Edit covar_redo.sh in CHATS/scripts to specify desired start and stop times.
Run it in batch mode:

>batch
at> ./covar_redo.sh
at> D

See if running:
>at -l
should print single line about batch job running