HP 3PAR StoreServ 7450 4-node HP 3PAR OS 3.1.2 Release Notes - Page 17
Persistent Ports Limitations, SLES11.1 Host with Device Mapper Multipathing
View all HP 3PAR StoreServ 7450 4-node manuals
Add to My Manuals
Save this manual to your list of manuals |
Page 17 highlights
Issue ID 81222 81383 81507 81667 81772 (also 75517) 83492 Item Description The only impact this has is the suspension of periodic resyncs for the duration of the online upgrade. eventlog: Critical Cluster thermal shutdown seen during soft reset of DCN1 cages After a soft reset of the DCN1 cage's controllers, we noticed the following events being generated. 2012-10-26 09:27:41 PDT 0 Critical Cluster thermal shutdown hw_node:0 Node 0 We have not noticed any issues caused by the generation of this event. A Newly Converted/Imported TPVV A volume that has been thinly imported via Peer Motion or converted should have Zero Detect on by to a thin volume using online conversion will not have the zero detect Default policy set by default. This does not affect the conversion process, which will perform zero detection correctly. The work around for this is to manually set the zero detect policy after the import or conversion has completed. System jammed with max limit of VV objects when running RC... In an RC setup with periodic sync, it is possible for the HP 3PAR OS to hit the limit of VV objects that can be created. When that happens, the periodic sync may not complete and the sync cycle may be missed. SP-4.1.0.-60: The simple upgrade from 3.1.1 + Patches to 3.1.2 wasn't successful Check Health sometimes incorrectly Due to an inconsistency in how PCI slot information is returned, reports matching empty slots checkhealth may display incorrect information about PCI cards being incorrectly matched between nodes. Example of erroneous output: Checking node Component Description Qty Node PCI card model differs for slot in node pair 2 Component -Identifier Description Node node:0 PCI card in Slot:4 is empty, but is not empty in Node:1 Node node:1 PCI card in Slot:4 is empty, but is not empty in Node:0 IO timeout with HP-UX 11i v2 Certain IO requests issued by HP-UX 11i v2 hosts have strict timeouts associated with them. If the HP 3PAR array to which the host is issuing IOs experiences a node reboot, the outstanding IOs may experience stalls while the node recovers. If the stall exceeds the timeouts, IOs can be marked as failed. This is a temporary condition. If the IO requests are tried again after the node recovery, they will succeed. HP recommends that a supported multipathing solution be installed and configured on the HP-UX 11i v2 host. This will automatically retry the IOs that timeout. Persistent Ports Limitations The goal of the persistent ports feature is to quickly transfer the WWN of one Inserv HBA port to another. Done quickly enough, the host multipathing software should not mark any paths as offline/unusable. However, with the current 3.1.2 release, the transition is sometimes not quick enough. The result is that, in some scenarios, the host multipathing software does notice the transition. In addition, in one scenario, the transition is actually too fast for the host to handle it correctly. The following scenarios have been observed during testing: • SLES11.1 Host with Device Mapper Multipathing Periodically, there will be multipath errors during failover or failback. These errors describe devices going offline and within seconds the devices are back on line. This is an intermittent issue seen about 10% of the time. Known Issues with the OS 17