sadly after several weeks of fault-free running we have had frame writer issues starting late Tuesday night. Only h1fw0 and h1fw1 are showing the problems, with h1fw2 stable. All three are running RCG2.3 code, the only difference is that fw2 has a local (non-QFS) disk. Some of the error messages from fw0 and fw1 suggest disk access issues.

One quick thing to try (which has helped in the past) is to power cycle the Solaris QFS-NFS servers. Under WP6179 we performed the following:

• reconfigured nds0,1 to serve h1fw1 frames
• power down h1fw0
• power cycle h1ldasgw0
• power up h1fw0
• reconfigure nds0,1 to server h1fw1 frames
• power down h1fw1
• power cycle h1ldasgw1
• while waiting for h1ldasgw1 to boot, h1fw0 restarted causing loss of frames
• power up h1fw1