hi, i have one EVA 3000 (stand-alone) an 2 EVA 4400 replicated. On july 28, primary Eva 4400 stop responding in CV . i logged a call to HP, sent logs, but near a weeks later i always have my EVA unreachable... LUN's works fine, non impact on production now, but i cant manage vdisk (nor in CV, nor in RSM...). Before this problem, i made a lot of trie to reinstall rsm 4 and change some zoning in san switches. HP think about a zoning problem (i checked and saw nothing wrong). Ok HP, it's good to "think about a problem" but i think their jobis to resolve, no justy to think ;) So if someone has ideas to help me i take !
Hi, can you share the logs here pls? there are some things which can be done, e.g.: 1. restart the CVE service and try to rediscover 2. reboot one of the affected EVA controllers to resync (if the affected EVA completely disappeared from the CVE, then it can be done only onsite), to return to the managable state
hi, of course i tried to reboot CV services, and reboot the 2 manap servers too. Now i can't see the primary EVA 4400 in any CV.
I also think i'll have to works on controllers, but i would not to do this without agree of HP Support...
My big problem is : "what are they doing ???" ok it's hollydays, ok my servers works, but no clones, no snaps, no new vdisks, etc... (and no replication i guess...)
If you have an EVA4400, you'll have Command View 8. Check in the Command View folder (C:\Program Files\Hewlett-Packard\Sanworks\Element Manager for StorageWorks HSV) for cvutil.exe.
Run cvutil.exe release_locks -wwn
Where is the WWN of the EVA that Command View cannot manage.
You can also check that the zoning is correct running cvutil.exe paths -wwn
Hi, no luck, i think my replica is broken cause all DR groups are initialized... #:(
Cvutils returns nothing interesting :
C:\Program Files\Hewlett-Packard\Sanworks\Element Manager for StorageWorks HSV>cvutil release_locks -wwn 5001-4380-013B-FAA0 Releasing SCMI lock on array 5001-4380-013B-FAA0 This command is for field service only! Are you sure (yes/no)? yes
C:\Program Files\Hewlett-Packard\Sanworks\Element Manager for StorageWorks HSV>cvutil paths -wwn 5001-4380-013B-FAA0 Status . done.
Path Vendor Product Firmware Node Serial SCMI Master WWN Number State ---- ------ ------- -------- ---- ------ ----- ------
C:\Program Files\Hewlett-Packard\Sanworks\Element Manager for StorageWorks HSV>
HP Support waked up, they asked to assign an ip address to controllers and try to take control of them. Also to set aptpolicy to " Port Based Routing Policy option 1" onto the san switches. If no successful : - stop the controllers (full stop) - stop manaps - reboot switches - restart EVA - restart manap - launch command View EVA
If no succes - send new Hps reports from manap - send new CVUTIL ouput
Hi, i tried this week to connect to the HSV 300 controler, i plug an ethernet wire, a PC, one manap on same subnet (192.168.0.x/24). The HSV controler said that all works fine !!! No event, no error in logs... I did also one thing asked by HP support but that cause a problem and i don't understand why : they asked me for modify san switches setup and set aptpolicy to 1. That command need to disable the swith, so i disabled one swith, set aptpolicy to 1, and enable switch quickly. All seems to be good, so i did the same on another switch, and so on (4 switches), waiting a little between each. And some minutes later, phone begun to RING !!! We lost some LUNS !!! Some virtual windows servers with DB reboot, other virtual servers saw nothing, som AIX server lost SOME DB, but not all... Very strange affair... May be a problme of latency when switch as enable ? Why ? All servers have 2 HBA cards, each on 2 san switches, etc... I thinked to disable or stop a switch in the chain would be transparent, NO ??? Do you have idea for why i got this kind of problem ? And also can i try something "soft" on the HSV controler to make the EVA seen by other equipements ? I would avoid to stop servers, eva, and switches like HP support asked to me... i works in hospital and stop everything is VERY hard...
i take a look at zonig configuration and launch access matrix tool onto the 2 fabrics (4 switches , 2 fabrics), how should i interpret this ? (see attached file)
We've seen this at multiple customers now with EVA 4000, 6000, 8000 and 4400. Every level of CV experiences the same issue and we have tried all versions of MPIO. HP support have not been able to help and have also pointed us towards zoning conflicts (which is not the issue of course).
The only solution is to bounce the EVA controllers which is not really that acceptable in a 24/7 environment.
Be interested if HP ever post this as a known issue or if you found any other solution.
Hi, we're just in validation phase. HP and integrator team changed controllers to disks cabling system (kown problem with some kind of cables) and reconnect continuous access links. HP said new EVA4400 fp ports need to be connected on same fp port number onto the other EVA, and not "cross connection" like EVA (as 4000/5000/ etc...) ... Now we're waiting but i think we'll don't have any problem (but not sure...).
I see i'm not the only one having major issues with the 4400's. We running 4 and HP recommended upgrading to 09501000 a few of weeks back.
Now one has completely failed. It appears as though the controllers have gone out of sync and neither is taking mastership. Thankfully its not quite in production.
I've also heard the problem could be temperature related in that the 1TB disks don't like it being too cold.
Is there any news on when the new code is getting released?
