Title: | *OLD* ALL-IN-1 (tm) Support Conference |
Notice: | Closed - See Note 4331.l to move to IOSG::ALL-IN-1 |
Moderator: | IOSG::PYE |
Created: | Thu Jan 30 1992 |
Last Modified: | Tue Jan 23 1996 |
Last Successful Update: | Fri Jun 06 1997 |
Number of topics: | 4343 |
Total number of notes: | 18308 |
A customer has just upgraded to V3.0 and is having severe problems with his file cabinet servers. Background - The customer has 2 nodes an 8700, and a 7000. The 7000 contains all of the users but the 8700 contains their old file cabinets which they wish to access using DSO. Configuration - Multi-lingual Multilingual ALL-IN-1 V3.0-1 ALL-IN-1 V3.0 (downgraded from V3.0-1 because of VMS 5.5-2 VMS 5.5-1 a mail problem) Users <----------DSO-------> Users old drawers Server limited to 50 client connects Symptoms - The File cab server on the 7000 starts looping, the file cab server on the 8700 is idle. At this point users either hang if they have a remote connection or slow down due to the high priority of the 7000 server hogging CPU. This problem can be cleared only by stopping and restarting each of the servers. It has been occuring about every 30 mins during the day. Errors - The logs contains lots of errors on each machine. On the 7000 they are mainly; %OAFC-E-INTERR, Internal error in File Cabinet Server Message: FCS has access violated and less frequently; %OAFC-E-THREADACCVIO, An FCS thread access violated, please submit an SPR. On the remote 8700 node the errors contain several; %LIB-F-INSVIRMEM, insufficient virtual memory Action - We attempted upping the process quotas for the server on the 8700 above the defaults for 50 users (eventhough we have limited remote and hence all connections to 50 as there are no local users). Tomorrow I am going to attempt increasing PGFLQUOTA even higher and see if the situation improves. I have not had very long to speak to the customer today so havent had time to monitor the server using the SRI option. My feeling is that it is the remote 8700 server which is access violating possibly caused by the memory error (they have also had insufficient dynamic memory), this seems to cause the local 7000 server to loop. I am also concerned about the mix of V3.0 and V3.0-1 servers - should this be ok ? Any other ideas or important areas to check ? Thanks Aston UK CSC
T.R | Title | User | Personal Name | Date | Lines |
---|---|---|---|---|---|
2557.1 | Further info ... | WAYOUT::CLARKE | Wed Apr 14 1993 12:41 | 12 | |
I have since found out that the users on the 7000 are using DSO to soley refile or delete documents from the remote drawers. My colleague pointed out to me that this was reported in a THR which was fixed in V3.0-1 (multiple copies and rfd's causing access violation and insufficient memory errors on remote server). We have advised that the customer should upgrade each server to V3.0-1 unfortunately this is being held up by a mail problem introduced by V3.0-1 FCS. The customer has already downgraded from V3.0-1 to avoid this problem. I will update this note if this doesn't resolve the acc vio problem. Aston | |||||
2557.2 | Sounds like you found a fixed bug :-) | CHRLIE::HUSTON | Wed Apr 14 1993 19:58 | 29 | |
Sounds like a nice problem :-) First off, the FCS that is access violating is the one that is writing it to the log file, always. There was a massive memory leak that was fixed in V3.0-1, it has to do with remote copy/refile of a number of documents (like doing a RFF or copy folder to remote drawer) there basically was a (65K * (n-1)) + n * 2K byte memory leak where n is the number of documents copied/refiled. This sounds like you found it! It also resulted in the server who did not run out of memory to acc vio most of the time, sound familar? You have two options: 1) Update both servers to V3.0-1 2) Wait for a PFR that will include all V3.0-1 fixes plus more. To avoid the mail problems that are introduced in V3.0-1, before you move objects from the INBOX, move them to another folder IN THE SAME DRAWER, then move from that folder to the other drawer. Or conversly don't do cross-drawer moves to the inbox. The problem basically is that the FCS cannot see the INBOX. Those are you two choices. I also believe there MIGHT be an ICF available to fix the INBOX problem, if so, you can update to V3.0-1 then install the ICF and go happily on your way. --Bob | |||||
2557.3 | Problem cleared | WAYOUT::CLARKE | Fri Apr 16 1993 13:40 | 8 | |
ICF #009 and ALL-IN-1 V3.0-1 installed now on the server machine. Mail problem and access violations have cleared. Customer is impressed with the turnaround on this problem. Thanks to everybody involved. Aston |