Hi all. I have what I hope is a simple question.

I have a Linux (vanilla 2.4.27) disk server which has several JFS 
partitions of about 200GB each. One of these partitions filled up, while a 
number of NFS clients were still trying to write to it. The following 
syslog entries started appearing:

Oct 22 17:17:38 swigdog10 kernel: jfs_create: dtInsert returned -28
Oct 22 17:17:38 swigdog10 kernel: txAbort: tid:49 dirty:0x0
Oct 22 17:17:39 swigdog10 kernel: jfs_create: dtInsert returned -28
Oct 22 17:17:39 swigdog10 kernel: txAbort: tid:264 dirty:0x0
Oct 22 17:17:51 swigdog10 kernel: jfs_create: dtInsert returned -28
Oct 22 17:17:51 swigdog10 kernel: txAbort: tid:272 dirty:0x0

as well as 

Oct 22 16:25:09 swigdog10 kernel: add_index: xtInsert failed!

repeated many times.

Are these messages simply the result of the filesystem being full? 
Wouldn't one expect to see a more traditional "no space left on device" 
sort of message instead? Or does this indicate something more dire? I ran 
jfs_fsck -f on the affected filesystem and it found no errors.

Any information you could provide would be greatly appreciated.

Thanks!

Raymond Prach
T.J. Watson Research Center
IBM - Yorktown
_______________________________________________
Jfs-discussion mailing list
[EMAIL PROTECTED]
http://www-124.ibm.com/developerworks/oss/mailman/listinfo/jfs-discussion

Reply via email to