info-vax@ucbvax.ARPA (09/13/84)
From: Provan@LLL-MFE.ARPA excuse me for repeating this question, but i still haven't solved the problem, so i'm hoping to get a wider distribution now that school's back in session. we have an ra80 and an ra81 hooked up to our vax-750 via your normal UDA50. we're running VMS 3.4. we're having a problem, and i want to find out if it's normal and whether or not we can do anything about it. after so many files are created on the RA81, when anyone tries to create another file, they get the following message: %CREATE-E-OPENOUT, error opening <file name> as output -RMS-E-CRE, ACP file create failed -SYSTEM-F-CTRLERR, fatal controller error it seems to have hit a maximum file count, since each file deleted seems to allow another to be created. when i initialized the RA81, i used the default maximum_files by not specifying the switch. according to my calculations, this default should be on the order of 100,000 files and "Show Device DUA1/Full" confirms this by displaying a "max. files allowed" of 111,384. i did specify the /Headers switch and gave a value of 5,000, which seems to be "close" (within a few hundred) to the number of files actually on the disk when the problem comes up. i'm under the impression that this value will only increase effeciency, but it's uncomfortably close to the magic number that's giving us trouble. we've got over 600,000 blocks free on the the '81, so that's not a problem. the problem happens to everyone on the system (all 3 of us) at the same time and in any directory we try. none of the directories being used are overly large. and to be complete: files can be deleted from any directory to allow anyone to create additional files. we also had this problem with our RA80, so i find it hard to believe it hasn't happened to anyone else. it vanished on the '80, probably just because we deleted enough files so we didn't hit the brink again. has anyone seen this or have any ideas what could be causing it? does it REALLY sound like a controller error? should i consider getting our hardware people to replace the controller to see if it goes away? i'm hard pressed to imagine how this could be a hardware problem, but i'm getting desparate. i really DO want to use those 600,000 blocks, but i'm finding it difficult under the circumstances. thanks to all who responded the first time. some of you may notice your input in this statement of the problem.