| Title: | DEC TCP/IP Services for OpenVMS |
| Notice: | Note 2-SSB Kits, 3-FT Kits, 4-Patch Info, 7-QAR System |
| Moderator: | ucxaxp.ucx.lkg.dec.com::TIBBERT |
| Created: | Thu Nov 17 1994 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 5568 |
| Total number of notes: | 21492 |
The customer is running:
Digital TCP/IP Services for OpenVMS Alpha Version V4.1 - ECO Level 2
on a AlphaServer 4100 5/400 4MB running OpenVMS V6.2-1H3 on a 4100
cluster (512MB each).
After 3/4 days of uptime the copies via FTP do not work anymore. The
error is "insufficient dynamic memory". If we shutdown UCX and start
again the problem disappears. We have increased already the number
of small and large buffers, but with no success.
Communication Configuration
Local host: l114dec05 Domain:
Cluster timer: 0
Interfaces: 0 Type: Default
Device_sockets: 0
Routes: 0
Services: 0
Proxies: 0
Free Maximum Minimum
Large buffers 0 250 50
Small buffers 0 1000 250
IRPs 0 0
Non UCX buffers 0
Remote Terminal
Large buffers: 0
UCBs: 0
Virtual term: disabled
System Memory Resources on 6-FEB-1997 09:27:15.92
Nonpaged Dynamic Memory (Lists + Variable)
Current Size (bytes) 79994880 Current Size (pagelets) 156240
Initial Size (NPAGEDYN) 79994880 Initial Size (pagelets) 156240
Maximum Size (NPAGEVIR) 319995904 Maximum Size (pagelets) 624992
Free Space (bytes) 65849024 Space in Use (bytes) 14145856
Size of Largest Block 64997504 Size of Smallest Block 64
Number of Free Blocks 4658 Free Blocks LEQU 64 Bytes 1530
Free Blocks on Lookasides 1766 Lookaside Space (bytes) 445632
Paged Dynamic Memory
Current Size (PAGEDYN) 4997120 Current Size (pagelets) 9760
Free Space (bytes) 3410864 Space in Use (bytes) 1586256
Size of Largest Block 3407632 Size of Smallest Block 16
Number of Free Blocks 66 Free Blocks LEQU 64 Bytes 63
What can we do to solve this problem?
Thanks in advance,
RUI
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 5202.1 | questions | UTRTSC::KNOPPERS | Oswald Knoppers | Thu Feb 06 1997 08:26 | 7 |
So where do you get this error? Is the problem with outgoing or incoming FTP. Do you get any opcom messages? Did you have a look at the FTP logfiles? We really need much more info. Oswald | |||||
| 5202.2 | Outgoing FTP | LISALF::BRITO | Fri Feb 07 1997 06:13 | 27 | |
The error appears in the log files of the command procedures (batch
run) that put the files in other systems (1000 files a day).
I am including an extract of the log:
.
.
.
$ SET NOVERIFY
$ @AMA$DISK:[FACT.BG.TMP]FTP_BG0_AMA_01_961121.COM
$ ON ERROR THEN CONTINUE
$ FTP bghp03
%FTP-E-NETERR, I/O error on network device
-SYSTEM-F-INSFMEM, insufficient dynamic memory
%DCL-W-SKPDAT, image data (records not beginning with "$") ignored
.
.
.
I have checked the operator.log and there was no ftp or dynamic
memory error messages.
I did not find any relevant information in the UCX$FTPD.LOG
files.
RUI
| |||||
| 5202.3 | not sure if this is an ucx problem | UTRTSC::KNOPPERS | Oswald Knoppers | Fri Feb 07 1997 06:47 | 11 |
Ok, can you do a 'ucx sho comm' and a 'ucx sho comm/mem' on this system? Another question. Is this one batch job running all the time and doing the 1000 ftp's a day? If so you might have a look at the accounting record of such a job. For instance there has been a bug in VMS that if you repeatedly spawn subprocesses one would loose pagefilequota. Maybe something like that is happening. Regards, Oswald | |||||
| 5202.4 | LISALF::BRITO | Fri Feb 07 1997 09:46 | 93 | ||
The cluster was rebooted today, so we can have a quiet weekend.
There is not one job. There are many jobs that run during the day. We
already checked the accounting. No hints from there.
Here are the result of the commands:
Communication Parameters
Local host: l114dec05 Domain:
Cluster timer: 5
Maximum Current Peak
Interfaces 20 3 3
Device_sockets 300 10 14
Routes 65535 3 3
Services 200 4 5
Proxies 20
Type: Ethernet Free Maximum Max Bytes Minimum Min Bytes
Large buffers 68 238 449344 34 64192
Small buffers 381 889 227584 127 32512
IRPs 20 200
Non UCX buffers 10
Remote Terminal
Large buffers: 10
UCBs: 4
Virtual term: disabled
MBUF Summary
Small_static Large_static Small_dynamic Large_dynamic
Total buffers 127 34 0 0
Free 87 33 0 0
Busy
Data 0 1 0 0
Header 7 0 0 0
Socket 12 0 0 0
Prot. control 12 0 0 0
Route 3 0 0 0
Socket name 0 0 0 0
Socket options 0 0 0 0
Fragment reassembly 0 0 0 0
IP address 3 0 0 0
Size of cluster 32768 64448 0 0
Free Current Peak Waits Drops
Small Buffers 40 45 0 0
Large Buffers 1 84 0 0
IRPs 3 8 11 0 0
Small clusters Large clusters Non UCX buffers
Free 0 2 0
We thought it might be a ucx problem, since the UCX shutdown and
restart solved the problem. But we do not exclude the possibility of a
problem like the one you mentioned. We already increased the account
quotas.
Username: STF_FACT Owner: STF-Factura��o
Account: FACT UIC: [121,2]
([TLP,STF_FACT])
CLI: DCL Tables: DCLTABLES
Default: FACT$DISK:[FACT]
LGICMD: LOGIN
Flags:
Primary days: Mon Tue Wed Thu Fri
Secondary days: Sat Sun
No access restrictions
Expiration: (none) Pwdminimum: 6 Login Fails: 0
Pwdlifetime: 90 00:00 Pwdchange: 31-JAN-1997 16:59
Last Login: 7-FEB-1997 14:32 (interactive), 7-FEB-1997 14:38
(non-interactive)
Maxjobs: 0 Fillm: 100 Bytlm: 65000
Maxacctjobs: 0 Shrfillm: 0 Pbytlm: 0
Maxdetach: 0 BIOlm: 200 JTquota: 8192
Prclm: 10 DIOlm: 200 WSdef: 8000
Prio: 4 ASTlm: 750 WSquo: 16000
Queprio: 0 TQElm: 100 WSextent: 49152
CPU: (none) Enqlm: 6000 Pgflquo: 400000
Authorized Privileges:
ALTPRI GRPNAM NETMBX SYSLCK SYSNAM TMPMBX
Default Privileges:
GRPNAM NETMBX SYSNAM TMPMBX
Identifier Value Attributes
PSI$X25_USER %X80010012
Regards,
RUI
| |||||
| 5202.5 | UTRTSC::KNOPPERS | Oswald Knoppers | Mon Feb 10 1997 04:06 | 5 | |
> The cluster was rebooted today, so we can have a quiet weekend. I am affraid we need these counters after the problem has happened. Oswald | |||||
| 5202.6 | UCXAXP::GEMIGNANI | Tue Feb 11 1997 19:11 | 14 | ||
Included in the V4.1 kit is an SDA extension for UCX.
It has been enhanced several times and we are sending the latest image
with the ECOs, as we publish them.
Can you try the following commands when in ANALYZE/SYSTEM:
SDA> UCX SHOW COMMUNICATION
SDA> UCX SHOW COMMUNICATION /MEMORY /FULL /ALL
The first should give you a good summary. The second will give you
details of the usage of each MBUF. You are most interested in those
marked as "Free".
| |||||
| 5202.7 | LISA2::BRITO | Thu Feb 20 1997 09:45 | 8 | ||
Thank you for your hints. The situation didn't repeat anymore (so far).
Anyway we added a batch job to the system to monitor the ucx memory.
If the situation repeats I will use the ana/sys commands.
Regards,
RUI
| |||||