I have immediate backups going to an adv_file device (on a SAN array), with a staging policy to move anything older than 13 days to tape devices. However, 2 weeks ago there was a server crash during a weekend full backup. Now that it is time to stage save sets from that time frame, I'm getting the following errors in daemon.log:
03/15/06 15:27:47 ansrd: ansrd_clone FAILED: errnum is 15002 and errstr is cannot open /emc1/38/45/004cdfde-00000006-37f42bd7-43f42bd7-00ca0000-0a0a0a0a file: No such file or directory
nsrstage: SYSTEM error: cannot open /emc1/38/45/004cdfde-00000006-37f42bd7-43f42bd7-00ca0000-0a0a0a0a file: No such file or directory
03/15/06 16:28:02 nsrd: nonexistant cloneid for ssid
And the stage completely fails to backup. Obviously I have some corrupt save sets that I need to delete, but how can I check for corrupted ssids and only delete those?
Run the mminfo to list you all the ssids on that volume. What you are looking in the report is ssid state, number of copies, clone id, time when it was saved (so all those older that 13 days are to be inspected). If you won't find anything wrong in such output you will probably need to manually stage ssid by ssid from that output until you find your naughty ssid.
However, there is another trick I believe how to play with this. You see, even it doesn't look so, what is says in that output IS your ssid Check this out:
C:\>mminfo -av -r "volume,ssid(60)"
volume ssid
hcrvelin.001 3f66515f-00000006-00181cd0-44181cd0-00010000-c0a8010e
hcrvelin.001.RO 3f66515f-00000006-00181cd0-44181cd0-00010000-c0a8010e
C:\>mminfo -av -q ssid=3f66515f-00000006-00181cd0-44181cd0-00010000-c0a8010e -r ssid,name,client
ssid name client
1580240 C:\tmp hcrvelin
C:\Appl\filedev>dir /s 3f66*
Volume in drive C has no label.
Volume Serial Number is FCF1-06A8
Directory of C:\Appl\filedev\90\00
15/03/2006 14:55 5,232 3f66515f-00000006-00181cd0-44181cd0-00010000-c0a8010e
1 File(s) 5,232 bytes
Directory of C:\Appl\filedev\90\00\notes
15/03/2006 14:55 176 3f66515f-00000006-00181cd0-44181cd0-00010000-c0a8010e
1 File(s) 176 bytes
Total Files Listed:
2 File(s) 5,408 bytes
0 Dir(s) 33,287,397,376 bytes free
So, the ssid as such as in media database is no longer at its place. With above steps you should be able to find it and fix it in not time.
Since there does seem to be some corruption at the SSID level maybe by adding another flag in your -q option for valid should weed out the suspect SSID and then you can work on staging off the SSID from this time frame. As noted earlier do report more fields for the output to really take a look at what is causing the failure. -r ssflags, state should help you out to id any problems.
ble1
6 Operator
•
14354 Posts
•
56186 Points
243
0
Posted March 15th, 2006 22:00
However, there is another trick I believe how to play with this. You see, even it doesn't look so, what is says in that output IS your ssid
So, the ssid as such as in media database is no longer at its place. With above steps you should be able to find it and fix it in not time.