topic Re: Error Disk full or maybe damaged in SAS Enterprise Guide

Error Disk full or maybe damaged

M_A_C — Thu, 04 Apr 2013 12:16:05 GMT

Hi,

So i always seem to get this error of disk full or maybe damaged which means that in the server i have no longer free space. I do the purging and delete everything i no longer use.

Is there a way to bypass that? (I know i can just ask the administrator to buy new servers but that would not be recommended)

Re: Error Disk full or maybe damaged

DBailey — Thu, 04 Apr 2013 13:23:16 GMT

Are you asking how you can bypass the error that tells you there is no more disk space available? Don't think that is feasible. If your system administrators are enforcing disk quotas, then you might could increase your allotment without purchasing new servers.

Re: Error Disk full or maybe damaged

M_A_C — Thu, 04 Apr 2013 14:22:28 GMT

Isn't there like a way to save the temporary tables on your own hard disk ?

Re: Error Disk full or maybe damaged

Astounding — Thu, 04 Apr 2013 14:30:55 GMT

Some of the answers depend on your operating system. For example, on Unix, a failed SAS job may preserve its WORK area. It would be worth looking for WORK folders that are just hanging around and should be deleted.

Are you using COMPRESS=YES or COMPRESS=BINARY? Depending on characteristics of your data, you might save a lot of disk space (at a small cost in CPU time to uncompress data sets as they are needed).

Do the WORK areas compete with permanently saved data sets for the same disk space? You might have old, permanent SAS data sets that are no longer needed. Is that what you mean when you save you purge/delete everything that you no longer need? If so, remember that you can delete WORK data sets that you no longer need in the middle of your program:

proc datasets library=work;

delete data sets I no longer need;

run;

quit;

Hope this helps.

Re: Error Disk full or maybe damaged

Doc_Duke — Thu, 04 Apr 2013 17:46:56 GMT

M.A.C.,

Seeing your other post, you have a huge incoming dataset. Judicious use of the KEEP statement or option will significantly reduce work data set size; EGuide does that as well as using VIEWs when you are using the task interface.

In a shared server environment, putting the WORK space on a separate spindle can do marvels. It can also be treated as a separate drive and not backed up.

Also in a shared server environment, the problem may be someone else's and you may be "collateral damage". The network admin can tell you if others are using all the space. As Astounding said, the server environment can make a big difference; maybe you could tell us more about your work setup.

Doc Muhlbaier

Duke