Article View: bionet.software.acedb
Article #2834Write locks in Acedb
From: Henry J. Cobb
Date: Thu, 23 Sep 1993 10:59
Date: Thu, 23 Sep 1993 10:59
14 lines
1091 bytes
1091 bytes
The semaphore system in Acedb seems to be slightly misused, and anyway won't scale over NFS mounts. NFS guarantees only one atomic action (without the presence of lockd, which is not universal): File creation. So is there any objection to throwing out the Semaphores and other junk and replacing this with a simple: open( lockfile, O_RDWR | O_CREATE | O_EXCL) or fail in sessionClose()? -- Henry J. Cobb hcobb@fly2.berkeley.edu SFB Tyrant
Technical Details
Message-ID:
<HCOBB.93Sep23105946@fly2.berkeley.edu>Path:
arc1-ca.pugleaf.net!gmd.de!xlink.net!howland.reston.ans.net!agate!fly2.berkeley.edu!hcobbReferences:
<jlmccarthy-080993095310@slavsing.lbl.gov>Show Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: f6d40,d5cf6c0f3faa5ab0 X-Google-Attributes: gidf6d40,public X-Google-ArrivalTime: 1993-09-23 11:26:43 PST Path: arc1-ca.pugleaf.net!gmd.de!xlink.net!howland.reston.ans.net!agate!fly2.berkeley.edu!hcobb From: hcobb@fly2.berkeley.edu (Henry J. Cobb) Newsgroups: bionet.software.acedb Subject: Write locks in Acedb Followup-To: bionet.software.acedb Date: 23 Sep 93 10:59:46 Organization: University of California, Berkeley Lines: 14 Message-ID: <HCOBB.93Sep23105946@fly2.berkeley.edu> References: <jlmccarthy-080993095310@slavsing.lbl.gov> NNTP-Posting-Host: fly2.berkeley.edu