This happens with us on a continuous basis. We are on Listserv 14.5 and were told by support that we need more memory for the server. Unfortunately 14.5 on Windows won't use more than 2 Gb of RAM (according to support,) so we are awaiting an upgrade. We set the Listserv service to automatically restart on failure and that has seemed to keep Listserv available though it isn't a long term solution.
Melissa Kenny
IIS Administrator
Harvard Medical School IT
Center for Educational Technology
20 Overland Street 4th Floor
Boston, MA 02215
(617) 998-6684 (desk)
________________________________________
From: LISTSERV site administrators' forum [[log in to unmask]] On Behalf Of Curtis Stevens [[log in to unmask]]
Sent: Thursday, October 06, 2011 9:17 PM
To: [log in to unmask]
Subject: Re: [LSTSRV-L] Heap Error
This is a look at a typical failure
Load_index: V=625 T=984 size=546944k
*** FIO file cache now totals 128232k. A list of cached files follows. ***
File size Usage Flags File name
--------- ----- ----- ---------
128230k 1 C:\LISTSERV\LISTS\ATDE-SW-WDM\ATDE-SW-WDM.DBRINDEX
2k 2 U C:\LISTSERV\MAIN\DIGESTS.FILE
5 Oct 2011 00:02:06 ATDE-SW-WDM digest (HTML) is being distributed...
5 Oct 2011 00:02:06 Distributing mail ("ATDE-SW-WDM") from
owner-ATDE-SW-WDM@LISTS.WDC.COM...
5 Oct 2011 00:02:06 Mail posted via SMTP to [log in to unmask]
5 Oct 2011 00:02:06 Done - 1 outbound file (1 rcpt).
5 Oct 2011 00:02:06 Reindexing BP-NAS-ZERMATT-DL LOG11...
Load_index: V=0 T=0 size=320k
5 Oct 2011 00:02:07 Reindexing NPI-REVMTG-ADMIN-DL LOG11...
Load_index: V=32 T=47 size=16896k
5 Oct 2011 00:02:08 Reindexing NPI-REVMTG-US-MANAGEMENT-DL LOG11...
Load_index: V=31 T=46 size=18304k
5 Oct 2011 00:02:10 Reindexing NPI-REVMTG-ASIA-MANAGEMENT-DL LOG11...
Load_index: V=31 T=47 size=20992k
5 Oct 2011 00:02:12 Reindexing NPI-REVMTG-EXEC-STAFF-DL LOG11...
Load_index: V=31 T=31 size=18240k
5 Oct 2011 00:02:15 Reindexing BP-SW-WDPHOTOS-DL LOG11...
Load_index: V=0 T=0 size=768k
>>> Error 8 from HeapAlloc() for 379908678 bytes <<<
This is happening 2min after midnight. More often than not the error occurs
after updating the C:\LISTSERV\MAIN\DIGESTS.FILE. This is a variant on the
problem here. These are now happening once every couple days, always at the
same time. The switch you are describing may be the answer.
I am experimenting by taking all the log files that are >100MB and renaming
them... This should take them all out of the indexing equation.
-------------------------------------------------
Curtis E. Stevens
Director, Standards & Features Technology
3355 Michelson Dr. #100
Office: 1-1041
Irvine, Ca. 92612
Phone: 949-672-7933
Cell: 949-307-5050
E-Mail: [log in to unmask]
Remember, you may only be blamed for something if you are actually doing
something.
-----Original Message-----
From: LISTSERV site administrators' forum
[mailto:[log in to unmask]] On Behalf Of Eckstine, Nate
Sent: Thursday, October 06, 2011 3:05 PM
To: [log in to unmask]
Subject: Re: Heap Error
Would it be permission related? The only heap alloc error I ever received
was from overloading indexing at midnight. We turned off a pre index heap
memory switch
I can't remember the exact switch.
nate
-----Original Message-----
From: LISTSERV site administrators' forum
[mailto:[log in to unmask]] On Behalf Of Curtis Stevens
Sent: Thursday, October 06, 2011 2:59 PM
To: [log in to unmask]
Subject: Re: Heap Error
Liam, thanks for your help! This list has 4 subscribers and about 800
messages totaling 11MB. The system has 4GB RAM and 2GB free disk space.
-------------------------------------------------
Curtis E. Stevens
Director, Standards & Features Technology
3355 Michelson Dr. #100
Office: 1-1041
Irvine, Ca. 92612
Phone: 949-672-7933
Cell: 949-307-5050
E-Mail: [log in to unmask]
Remember, you may only be blamed for something if you are actually doing
something.
-----Original Message-----
From: Liam Kelly [mailto:[log in to unmask]]
Sent: Thursday, October 06, 2011 1:42 PM
To: Curtis Stevens
Cc: LISTSERV site administrators' forum
Subject: Re: Heap Error
> I am getting "Error 8 from HeapAlloc() for 379908678 bytes" during a
> reindexing operation. The list is average sized. What is this
> indicating?
It's a memory allocation failure. When you say the list is 'average sized',
what does that mean? What is the total size of the archive for this list,
and what is the single largest archive file for it?
In nearly all cases, a HeapAlloc failure is an actual memory or user quota
issue, but I've seen one case where it was caused by messages with corrupt
base64 data. If that's the case, you would see MIME parser errors in the
LISTSERV log immediately prior to the crash.
--
Liam Kelly
Senior Consulting Analyst
L-Soft international
[log in to unmask]
Interested in other L-Soft products? See what else we have to offer here
<http://www.lsoft.com/products/products.asp>
archives at <http://www.lsoft.com/news/newsletter-us.asp>
############################
To unsubscribe from the LSTSRV-L list:
write to: mailto:[log in to unmask]
or click the following link:
http://peach.ease.lsoft.com/scripts/wa.exe?SUBED1=LSTSRV-L&A=1
############################
To unsubscribe from the LSTSRV-L list:
write to: mailto:[log in to unmask]
or click the following link:
http://peach.ease.lsoft.com/scripts/wa.exe?SUBED1=LSTSRV-L&A=1
############################
To unsubscribe from the LSTSRV-L list:
write to: mailto:[log in to unmask]
or click the following link:
http://peach.ease.lsoft.com/scripts/wa.exe?SUBED1=LSTSRV-L&A=1
############################
To unsubscribe from the LSTSRV-L list:
write to: mailto:[log in to unmask]
or click the following link:
http://peach.ease.lsoft.com/scripts/wa.exe?SUBED1=LSTSRV-L&A=1
|
