Status for Andrew DeFaria

Cleaning up Multisite Packets

First order of business was to attempt to clean up multisite packets that reside in the shipping bays for both sons-clearcase and sons-sc-cc as much as possible. As per my prior work there seems to be huge sync packets to sync, which takes time. I wanted to attempt a chmaster on an older branch to see how that changes from sons-clearcase -> sons-sc.cc. Part of the chmaster involves informing the other replica of the change. This happens through the normal multisite syncreplica. If the bays are full of huge packets then I need to process them first. One problem I hit was running out of space on sons-sc-cc. Normally this is not a problem as there is enough space on the C drive where the vobs reside. But with these huge packets going back and forth I was running out of space. Cleaned up some space and attempt to import all packets on sons-sc-cc. I also attempted to scrub the cleartext pool on sons-clearcase, which has grown to 4 gig! The cleartext pool is a caching mechanism thus since Clearcase can reconstruct the cleartext pool at any time (cleartext is mutable) I figured I could save 4 gig.

Testing chmaster

Tested out that I cannot check out, and back in, and element on the rel_1.0 branch from a view on sons-sc-cc. I then attempted to transfer mastership of the rel_1.0 branch -> sons-sc-cc but received the following error:

[ccadmin] sons-clearcase:ct chmaster SantaClara brtype:rel_1.0@\\salira
cleartool: Error: Branch type "rel_1.0" has branches (with default mastership) that have outstanding checkouts.

Actually there are still checkout on the rel_1.0 branch in, for example, the view YXiu_view_desktop (e.g. salira/neopon/build/makefile).

Ran dbcheck on salira vob to fix corruption

10:40 Pm: Decided to give up on the testing of chmaster and get the vob fixed. Locked salira vob. Started copy of db

:10:43 Pm: Dtarted keybuild procedure. Keybuild failed with:

db_VISTA Version 3.20
Key File Build Utility
Copyright (C) 1985-1990 Raima Corporation, All Rights Reserved

initializing key file: vob_db.k01
initializing key file: vob_db.k02
initializing key file: vob_db.k03
initializing key file: vob_db.k04
processing data file: vob_db.d01, total records = 3555277
 record:       9000
 record:      19000
 record:      29000
 record:      39000
 record:      49000
 record:      59000
 record:      69000
 record:      79000
 record:      89000
 record:      99000
 record:     109000
 record:     119000
 record:     129000
 record:     139000
 record:     149000
 record:     159000

keybuild failed with an exit code of 58. Ran keybuild again... This seems to be going better... Did d01 file. Proceeded to work on the d02 file then (11:07 Pm):

record:   863000

*** db_VISTA database error -901 - system error

Bad read  863475
    863474processing data file vob_db.d02, total records = 1

 record:       1

key file rebuild completed

Hmmm... Doesn't seem like the key file rebuild was really completed. I wonder... Should I try again? Trying again...

Third times a charm they say! keybuild ran to completion but for a while it was touch and go as sons-clearcase was not responding. Now, however, I can import the packets that were stuck... Well most of them:

Applied sync. packet sync_SantaClara_26-Jun-06.02.00.01_5308 to VOB \\sons-clearcase\VOBs\salira.vbs
Multitool.exe: Error: Database identifier (dbid) not found in database: "\salira".
Multitool.exe: Error: Could not get oplog entry with order:2886884 from replica:
China with oplog_id:376595: reference to non-existent ClearCase object.
Multitool.exe: Error: Could not check oplog entry for divergence: reference to non-existent ClearCase object.
Multitool.exe: Error: Cannot apply sync. packet sync_China_26-Jun-06.16.32.42_3292_1 to VOB replica \\sons-clearcase\VOBs\salira.vbs: reference to non-existent ClearCase object

Damn. Ran syncreplica -import again and everything got processed. I'm glad it's processed but I can't help but wonder why I hit these errors...