Thunderstone Support Forums

I posted this on the wrong forum, so I'm re-posting it here.

I am unable to create an index on my database. See error below. There is plenty of disk space and no obvious sign of disk failure. I'm running True 64 Unix V4.0F. Texis 4.04.1073572878.

Thanks,
Scott

+ tsql -q set indexspace='/usr2/pub/rds/db/DBIndex'; set addindextmp='/usr3/tmp'; set addindextmp='/usr1/tmp'; set indexmem=60; set
addexp='\alnum{1,5}&=\alnum{1,5}'; set addexp='\alpha{2}:=\alpha+'; create metamorph inverted index xmBNI_all on BNI(Art_title
\Atitle\Author\Book_author\Book_title\Eabstract\Kw_idx\Srch_idx\Text\Titlesrch\Atype_term\Pubtype\Con_term\Geo_code\Geo_term\Graph_t
erm\Ind_term\Mark_term\Prod_code\Prod_term\Publisher\Trade_name, Date)
106 You may be low on disk space. in the function create index
005 Corrupt operation at 0x3D0FC9E in KDBF file /usr3/tmp/T02479i.dat: error 5: I/O error in the function kdbf_readchunk
015 Internal error: incomplete block write to KDBF file /usr2/pub/rds/db/DBIndex/xmBNI_all.dat in the function kdbf_flush
000 SQL failed

Check your system logs. There's not much Texis can do if the OS fails to do requested disk I/O's.

I've checked the system logs and see no indication of disk problems or failures. There were errors in the vortex.log indicating problems with the SYSLOCKS file. I restarted the machine stopped the monitor processes and deleted those files plus some temp tables that I found. I then retried with the same result. Any other suggestions?

Perhaps run a disk excersiser/test program on the /usr3/tmp filesystem to try to detect errors.

If the error only occurred after the index was progressing a while, it may be a bad sector in free space that is only used after the disk starts to fill. Try filling that filesystem by writing a temp file and see if you get an I/O error. Eg.:

cat /dev/zero > /usr3/tmp/bigfile

and then delete it. (You'll obviously get "file system full" or "no space left on device", but look for I/O errors.)