System going down from time to time

We have client who has been running Navision since version 1.20. The database is still the same but C/SIDE is today in version SE2.60D. From the beginning there has been problems with the server once or twice a year. The server goes crazy and the CPU peaks (It’s not CPUMONI that is causing the problem). The only solution has been to reeboot the server if possible. Sometimes the CPU Load is so hig that the OS doesn’t respond. Most of the times it has been possible to just start the server again. Other times the effect has been a corrupted database. The last time the database was corrupted during a batch post of a lot of invoices. The error was 19 in module 1185. After that we couldn’t acces table 36 and 37 in one of the companies. The primary keys was damaged. The system is running on NT4 SP6 wit 512 Mb RAM and the database is split in four parts over four mirrored disks in a RAID-1 Controller with Write Trough. The cache is set to 350Mb and the commit cache is enabled. There is approx. 40 users. Has enyone experienced something similar or has any ideas? //Lars

Lars - What other services or applications are running on the server? Also, how much disk space is available on the server. The problems we’ve seen that are similar, especially the CPU peaks, seem to be associated with the issues above. Also, if the database corrupted, they probably have an internal hardware cache running? Allen Beck Allen Beck President Beck Consulting Alameda, CA & Bellevue, WA allen@beckconsulting.com 800-456-8474

Hi Allen. No other applications is running on the server. There is approx. a total of 6Gb free diskspace on the disks and the Cache settings on the RAID Controller is set to Write Through. So there is no HW caching involved. One thing I came to think of is that we have intercomany transactions with both reading and writing in other companies. Is there anyone who has experienced these kind of problems together with extensive use of CHANGECOMPANY? The last time the database was corrupted the test of table 36 reported error in the primary key for that table with all blank values (also in companyname which in fact is a part of the key). //Lars

I’m not a technician but what I heard is that there’s a memory-leak problem related with NF 2.01 on NT-4. NF is running slower and slower untill the server is restarted. Funny enough it is not enough to restart the FinServer. You have to completely shut down the NT-4 server. ------- With best regards from Switzerland Marcus Fabian

While I think I have heard of such a memory leak we have some customers that hardly ever reboot their servers but maybe their databases are smaller and this problem only affects databases that are large or on multiple files for example. Lars says though that it is in 2.6D now. No problems with 2.6D customers with this problem. Craig Bradney Project Manager - Technical Navision Solutions & Services Deloitte Growth Solutions Deloitte Touche Tohmatsu P:+61-2-9322-7796 F:+61-2-9322-7502 E:craig_bradney@deloitte.com.au

We have the same problem in our database - we have to shut down from time to time (once a month/every second month). We are running a NF 2.0 on a NT 4 server. It is a 12 GB database, mirrored on 4 disks. I have tried to find out why this happen, but I haven’t succeded yet - maybe this is the answer? Mads Morre

Since this seams to be a more or less common problem, I’ve move this topic from the NF Tech forum to the NF Known Errors forum. Best regards, Erik P. Ernst, webmaster Navision Online User Group

Thank’s for Your replies. I think we have came up with a solution now. I had a theory that it was NT that swapped Navisions cache to disk and it seems as if I was right. After decreasing Navisions cache so that there is 250Mb RAM left for the OS the problems has stopped. My client today has 250Mb cache for 6Gb db and 40 users and the system has been running fine for several weeks now. I also see in the performance monitor that NT isn’t swapping any pages to disk. Strange that NT should need so much RAM to leave Navision’s cache alone. At least the problems have stopped now after this change. So: If You have problem with intermittent crashes or a CPU that goes wild. Check the performance monitor an look at the pages/sec. to see if NT is swapping memory. In that case this could be Your problem and maybe a workaround is to free more RAM to NT. //Lars