| Title: | ase |
| Moderator: | SMURF::GROSSO |
| Created: | Thu Jul 29 1993 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 2114 |
| Total number of notes: | 7347 |
Hi Presentation of the problem. My customer want to make some tests before its ASE cluster go in production. One test, which failed, consist to power off the 2 power supply of a single HSZ40. This HZS40 control the half of disks (JBOD) of the configuration and they are mirrored by LSM to the other half of disks which are controlled by a second HSZ40. Configuration details of the Cluster ASE. 2 systems configured with ASE rev 1.3 : UX03 Digital Unix version 3.2d-2 UX04 Digital Unix version 3.2d-2 2 single HSZ40 software rev 3.02 One HSZ40 on each bus between the systems UX03 and UX04 Alle disks are JBOD disks and are at the good revision for ASE and for the new software rev of the HSZ40 Only disks on these 2 shared bus, no tape or cdrom. ASE services : This cluster is configured with 2 disks services in an environment SAP SAPDB service : composed of 6 Advfs domains Placement policy=favor member UX03 SAPAS service : composed of 1 Advfs domain Placement policy=favor members UX04 Start and Stop scripts are the standards scripts of SAP-DIGITAL. all domains ADVFS are composed of one volume LSM with 2 Plexes MIRRORED ON THE 2 HSZ40. All patches knowed for LSM, ADVFS, Digital Unix and ASE are already installed on both systems. THE PROBLEM : For test purpose ,we power off one HSZ40 (plug out the 2 power cords). Results : _SAP application crash. _Not possible to use ASEMGR to have infos on the 2 services. _"unrecoverable errors" on some volumes under LSM. We have not waiting more as 10mn to see if the services can come up again QUESTIONS : Is this problem knowed ? Is this problem a true problem. ASE dont need more time to allow the services to come up again ? If you experience this problem or have a solution to resolve this problem, please let me know. Thank You in advance. Bertrand Lobstein [Posted by WWW Notes gateway]
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 2005.1 | KITCHE::schott | Eric R. Schott USG Product Management | Fri Apr 18 1997 12:56 | 10 | |
Hi Do you have error logs of the system at the time of the crash? Was all the data lsm mirrored? I suggest you run sys_check, to gather up more data...it is impossible to see what happened with such little data. http://www-unix.zk3.dec.com/tuning/tools/sys_check/sys_check.html | |||||
| 2005.2 | Answer to Eric R. Schott. | NNTPD::"bertrand.lobstein@lvg.mts.dec.com???@lxa506.lvg.dec.com" | lobstein bertrand | Tue Apr 22 1997 16:42 | 25 |
Hi Eric Thank You a lot for your answer. Sorry for my silence but these two weeks i am alone for unix in the service and i have a lot of work. > Was all the data lsm mirrored? Yes all data was lsm mirrored. > I suggest you to run sys_check ... I have already sys_check and with customer we schedule to run this tool next week or friday afternoon if i have no urgent work to do. By the way i ask the customer to send me the error logs on a floppy because i have only print out. Please Eric can you send me a unix node address where i can send you the error logs when i receive them from customer. Thank You a lot Eric. Bertrand. [Posted by WWW Notes gateway] | |||||
| 2005.3 | Not the official way !!! | BACHUS::DEVOS | Manu Devos NSIS Brussels 856-7539 | Tue Apr 22 1997 19:32 | 10 |
Bertrand,
Eric is a Product manager, and we all can thank him to pass its time
(early morning and late evening) to post replies to our problems in
the field, but he is NOT the guy to send daemon.log. There is official
channel for that. Give a call to TSC +32 2 729 77 44 for help in
Belgium or post your data as a reply to this note and maybe somebody
around the world will voluntary try to help you.
Manu.
| |||||