I am having a server abend frequently. Where do I start to figure out why? It is netware 3.15 and nlm 1.2 i think. Extremely LARGE files and HUGE volumes.

Dan,

You mean Netware 3.12?

Normally Novell will report what NLM caused the abend. However, using TTS on LH files can cause problems, as can using compression on volumes with a large amount of data in them.

Try getting detect.ncf from novell. It is a tool that is used in certification that can help finding the problem NLM. You also might try contacting Novell.

John Revelation

John,

Some background info on the server and the network:

-Proliant 800

Everything in this server is from Compaq.

-Ram Memory……….:400 MB

Also from Compaq.

-Disk Space divided

into 2 volumes:

- SYS: 500MB

- VOL1: 33+GB…….: 34 GB

-Novell version……: v3.12

Patches all current.

-Frame protocol……: v802.3

                   : Fast ethernet

                   : Cabling is less than a year old (call cat5).

-Typical workstation.: Win95a with 32mb RAM, Intel network cards,

                     P233mmx.  And yes, we're running AREV from

                     the workstation based on instructions from

                     this website.

-Main applications:

1. AREV …………: v3.12

  Unlimited vers.

2. AREV NLM………: v1.10

2. Arcserve ……..: v6.1.

3. Inoculan ……..: v4.0

4. Alexander SPK….: v2.50

All of the above are current vis-a-vis patches. Alexander SPK

is a Server Protection Kit from www.alexander.com which takes

over when the server crashes. The SPK identified AIO.NLM as the

culprit NLM in one of the more recent abends (it can't identify

the problem NLM in all cases eg. catastrophic crash, however, it

was able to capture all the server screens in one of the crashes).

We temporarily disabled AIO.NLM but the server still crashed.

-We are using default values in the Revelation NLM (ie. we are

not passing parameters to the program).

-We have thousands of files, the largest is about 3.6 GB

(that's a gigabyte, and that's one file - a REVxxxxx.* file.

We're aware of the 4gig limit).

-A typical business day would have from 70-100 users logged in.

-We've talked to Novell here in San Francisco. They recommended a

Novell auhorized vendor who has looked at the server. Based on

his recommendation (he is an MCNE, MCSE) and that of another

very knowlegeable consultant, we've changed practically everything

in the server. He's also managed to tune the server so that our

ECB count is now 0 and dirty cache is less than 100 (used to be in

the 4-5K range). The server still abends. The consensus at this

point appears to be that we are 90+% sure it's not the hardware.

Let me know if you need more info.

Thanks,

George

John,

Some background info on the server and the network:

-Proliant 800

Everything in this server is from Compaq.

-Ram Memory……….:400 MB

Also from Compaq.

-Disk Space divided

into 2 volumes:

- SYS: 500MB

- VOL1: 33+GB…….: 34 GB

-Novell version……: v3.12

Patches all current.

-Frame protocol……: v802.3

                   : Fast ethernet

                   : Cabling is less than a year old (call cat5).

-Typical workstation.: Win95a with 32mb RAM, Intel network cards,

                     P233mmx.  And yes, we're running AREV from

                     the workstation based on instructions from

                     this website.

-Main applications:

1. AREV …………: v3.12

  Unlimited vers.

2. AREV NLM………: v1.10

2. Arcserve ……..: v6.1.

3. Inoculan ……..: v4.0

4. Alexander SPK….: v2.50

All of the above are current vis-a-vis patches. Alexander SPK

is a Server Protection Kit from www.alexander.com which takes

over when the server crashes. The SPK identified AIO.NLM as the

culprit NLM in one of the more recent abends (it can't identify

the problem NLM in all cases eg. catastrophic crash, however, it

was able to capture all the server screens in one of the crashes).

We temporarily disabled AIO.NLM but the server still crashed.

-We are using default values in the Revelation NLM (ie. we are

not passing parameters to the program).

-We have thousands of files, the largest is about 3.6 GB

(that's a gigabyte, and that's one file - a REVxxxxx.* file.

We're aware of the 4gig limit).

-A typical business day would have from 70-100 users logged in.

-We've talked to Novell here in San Francisco. They recommended a

Novell auhorized vendor who has looked at the server. Based on

his recommendation (he is an MCNE, MCSE) and that of another

very knowlegeable consultant, we've changed practically everything

in the server. He's also managed to tune the server so that our

ECB count is now 0 and dirty cache is less than 100 (used to be in

the 4-5K range). The server still abends. The consensus at this

point appears to be that we are 90+% sure it's not the hardware.

Let me know if you need more info.

Thanks,

George

So is this it? Will no one go any further towards helping figure this problem out?

View this thread on the forum...