Sun System Handbook - ISO 3.4 June 2011 Internal/Partner Edition | |||
|
|
Solution Type Sun Alert Sure Solution 1001336.1 : Booting From a Sun StorEdge 6920 Release 3.0 (Running Baseline 3.0.0.25) May Fail With "Fast Data Access MMU Miss" Errors
PreviouslyPublishedAs 201798 Product Sun StorageTek 6920 System Bug Id <SUNBUG: 6306398> Date of Resolved Release 11-JAN-2006 Impact When attempting to boot a Solaris SPARC host from a Sun StorEdge 6920 array with System Processor (SP) version 3.0 (running baseline version 3.0.0.25), the boot process may fail intermittently. If this occurs, the Solaris boot will need to be retried. Contributing Factors This issue can occur on the following platform:
Note: Sun StorEdge 6920 arrays with System Processor (SP) version 2.0, running baseline version 2.x are not affected by this issue. The described issue only occurs on Sun StorEdge 6920 arrays where the host is booting from a LUN on the array. To determine the baseline version running on the array, do the following: 1. Log into the "Storage Automated Diagnostic Environment" web console as the "storage" user. Sun Storage Automated Diagnostic Environment -> Inventory -> Service Processor 2. Check the "Details" view for the version listed (e.g. 3.0.0.25). Note: If there is not an "Inventory" tab after selecting "Sun Storage Automated Diagnostic Environment", but there is a "Service" tab, then release 2.0 is running on the array and the system is not affected. Symptoms If the described issue occurs, the following "Fast Data Access MMU Miss" message will be seen on the host console: Rebooting with command: boot Boot device: rootdisk File and args: Fast Data Access MMU Miss {1} ok Note: There can be other causes of "Fast Data Access MMU Miss" messages during boot, and there is no method in a normal user environment to determine that the boot failure was definitely due to this issue. However, it would be a good indication that the Sun StorEdge 6920 array is affected by this issue if, for example, booting from the array was successful every time when running a baseline version of 2.x, but the booting started to fail with "Fast Data Access MMU Miss" messages only after upgrading the array to baseline version 3.0.0.25. Workaround If the described issue occurs, the system will return to the "ok" OBP prompt, and the Solaris boot will need to be retried. At the "ok" prompt, type "reset-all" (without the quotes) and press the enter key. If the OBP "auto-boot?" setting is "true" on the host, then the host will automatically reboot after resetting. If this OBP setting is "false" then the host will return to the "ok" prompt. At this point, that host's normal variation of the "boot" command should be used to boot the host. Note: The boot may fail a number of times. Apply the above workaround until the host successfully boots. Resolution This issue is resolved in the following release:
Note: Patch 118396-29 should only be applied as part of a full upgrade to baseline version 3.0.0.30, and not on its own. Baseline version 3.0.0.30 is available online by using the PatchPro functionality on the arrays Service Processor (SP), or by using a Sun StorEdge 6920 array version 3.0.0.30 patch CD available from your Service Provider. References<SUNPATCH: 118396-29>Previously Published As 102106 Internal Comments It is believed that this issue is fixed in DSP firmware patch 118396-29 (which must not be applied on its own, but only as part of a full SE6920 upgrade to Unity 3.0.0.30 using PatchPro online or with a Unity 3.0 Build 30 Patch CD). That patch has fixed similar behavior reproduced by PDE. However, the affected customer will not verify that the fix in the patch definitely resolves this issue. In addition to bug# 6306398 (where the SE6920 behavior did contribute to the boot failure), there is bug# 6322050 which is not an SE6920 issue, but can produce similar symptoms of intermittent boot failures. BugID 6322050 has its own workaround listed in the bug (set FC switch ports to be F-ports), while there is further investigation by PDE. If there is an issue booting from an SE6920 and the error message is not "Fast Data Access MMU Miss", it might be the result of bugID 6322050 as it seems to produce a variety of errors. To identify that the customer is running Release 3.0 baseline 3.0.0.25 from an SE6920 "Solution Extract" tar file, look in the file:
For example:
Internal Contributor/submitter sam.gibson@sun.com Internal Eng Business Unit Group NWS (Network Storage) Internal Eng Responsible Engineer sam.gibson@sun.com Internal Services Knowledge Engineer jeff.folla@sun.com Internal Escalation ID 1-10845464 Internal Resolution Patches 118396-29 Internal Sun Alert Kasp Legacy ID 102106 Internal Sun Alert & FAB Admin Info Critical Category: Availability ==> Regression Significant Change Date: 2006-01-11 Avoidance: Patch, Workaround Responsible Manager: Lorraine.Dilliway@Sun.COM Original Admin Info: [WF 11-Jan-2006, Jeff Folla: Sun was waiting to hear back from the customer on if the patch resolved the customer's issue. The customer never confirmed this but engineering is confident that the patch will solve this issue. Sent for release.] [WF 10-Jan-2006, Jeff Folla: Sent for review.] Product_uuid 67794720-356d-11d7-8ef2-ce2ac2bc9136|Sun StorageTek 6920 System ReferencesSUNPATCH:118396-29Attachments This solution has no attachment |
||||||||||||
|