I am using nginx for reverse proxy with caching. And I have a backend that sometimes returns 200 HTTP code for failed requests, with just an empty set for a body. I'd like to configure nginx in a way that would prevent storing of such responses.
I tried setting no_cache like this:
if ($upstream_http_content_length = "2") { set $nocache 1; } # empty response
proxy_no_cache $nocache;
But that doesn't work. What are my other options (besides fixing backend to behave properly)?
I have a few scripts I'm running that are sending an email. I wanted to configure everything to be able to use simple mail() functions and have emails coming from my gmail account, without exposing my server. However, all of the guides I've seen end up with postfix adding a line exposing the hostname of the local server and my public IP address.
How can I configure postfix so that it looks like the email was sent directly from gmail, without relaying through another host? And I'd like to configure it on postfix end, rather than having to rewrite the script to go to gmail directly. mail() is just so convenient!
I'm seeing these error messages in my SonicWall firewall (SonicOS Enhanced 6.2.7.1-23n):
Ethernet Header
Ether Type: IP(0x800), ......
IP Packet Header
IP Type: UDP(0x11), Src=..., Dst=...
UDP Packet Header
Src=[5060], Dst=[5060], Checksum=0x416c, Message Length=991 bytes
Application Header
Not Known:
Value:[1]
DROPPED, Drop Code: 702(Packet dropped - Policy drop), Module Id: 27(policy), (Ref.Id: _1857_rqnke{Ejgem) 4:3)
I've googled the heck out of all combinations, but I can't seem to find what this is. I see max of 404 code, and no explanation how to trace this drop to the policy. What's more mysterious, (this is SIP protocol) it allows the first exchange to establish the call, but it blocks any packets after that, like re-INVITE when timeout is approaching, or BYE when the other party hangs up.
Any pointers would be greatly appreciated!
I'm trying to monitor CPU usage of VMware host over SNMP, and I'm not getting the result I expect. I'm requesting this resource with snmpwalk: HOST-RESOURCES-MIB::hrProcessorLoad AKA hrProcessorTable. I get a list of CPU cores, and a number. I'm assuming % load. Averaging them out, I get 38% load. But when I compare to reported CPU load in vcenter, I get 65% load (on hosts list tab). I tried searching for OIDs on the web, but can't find anything more specific, so a lot is left for my interpretation.
What am I doing wrong? Is there a different OID I need to be looking at? Or am I not interpreting the number I get in SNMP correctly?
We have an application that came with MSDE 2000 version (yeah, quite a while ago) that is no longer supported, but we have not replaced it with something else. It works for us just fine, and functionality is OK, so we're not looking to replace it. But over the last 10 years, the amount of data in it accumulated to the point that it reached 2GB license limit for MSDE edition. We are required to keep a lot of this data, and since it grows exponentially, I'm not able to free up enough space to last for more than a few weeks to a month. The server is also running MS Windows 2003.
What is the most painless option to lift this license limit? Can I apply "Standard" license to this instance to lift these limitations? Do I need to migrate to a different distribution?
I have concerns with application working under newer distribution. But another concern is upgrade path is not easy: 2008 gives only 2 more G of database space, which is a shorter term fix than I'd like (may be a year or little more). 2014 runs only on server 2008 or higher, so I'd need to upgrade OS, or install it on a separate server (so another license is needed). So keeping it version 2000, and just upgrading the license would be the best bet, but I can't find any information on that. All google comes up with (rightfully so) is how to get rid of 2000, in favor of 2008.
How can I mitigate POODLE SSL vulnerability when using stunnel as HTTPS reverse proxy?
I am using heartbeat with MySQL master-master configuration, and using mon to check up on MySQL health. All works well for detecting downed host or broken MySQL. But the problems start after takeover, when replication collisions start piling up.
Is it possible to add a hook to heartbeat, to make it wait before taking over the IP address, by running an external script? I wrote one that checks up on replication, and exits with code 0 or 1 depending if all clear to proceed, and waits for replication to catch up completely to avoid collisions.
I've set up DRBD on 2 nodes, and started using it yesterday. After about an hour, it had resynced 50% of the partition. Another 12 hours passed, and it's up to 79%, and moving VERY slow.
Here's what cat /proc/drbd shows:
1: cs:SyncTarget ro:Primary/Secondary ds:Inconsistent/UpToDate C r-----
ns:464931976 nr:191087032 dw:656013660 dr:214780588 al:100703 bm:21100 lo:7 pe:0 ua:0 ap:7 ep:1 wo:f oos:92241852
[==============>.....] sync'ed: 79.2% (90076/431396)M
finish: 76:13:38 speed: 332 (8,680) want: 19,480 K/sec
I looked at network traffic, and I am using between 1M and 20M on 1G interface. Tried running iperf while all this is going on, and I get 930M reading. Tried adjusting syncer rate to 10M, 50M, 500M to no avail. Tweaked packet size without luck too.
Now, the caveat, as you can see from the status, is that my primary node is inconsistent. So I assume that OS is working with essentially a secondary node while resync is going. But given that throughput is so low, I don't understand why sync is not faster.
Any ideas on what I can try next? Estimated finish of 76 hours is not something I look forward to :( Especially not knowing the reason, so come an outage of sorts, I wouldn't know how to bring the array to consistency fast.
Thanks!
EDIT: I tried the following settings in the net section to no avail:
sndbuf-size 512k;
max-buffers 20480;
max-epoch-size 16384;
unplug-watermark 20480;
EDIT 2: For no apparent reason, speed jumped to 10~30M, after I stopped tweaking all the configs. Got up to 98.8% synced, and dropped back to ~300K. No messages in the logs on neither of the servers. Coincidentally, I see a spike in INSERT activity on MySQL database that runs off of this partition. Any ideas?
EDIT 3: Version: 8.4.2 (api:1/proto:86-101)
I seem to have a strange issue with my VMware cluster, where I get inconsistent SCP transfer rates. I have Cluster1 and Cluster2, physically in different regions. I need to transfer large vmdk from Cluster1 to Cluster2. Here is what I get:
Testing network bandwidth with iperf shows that I have 200-300Mbps between the locations consistently. Network connection in both clusters, as well as internet is gigabit.
What would cause VMware to limit CPU usage for compression, and what would limit it's transfer rates outside of the cluster?
PS: inside the cluster I am going through public IP, and guest VM is on a different host. So theoretically VMware shouldn't know that the transfer is somewhat local.
EDIT: Cluster1 is 4.1, Cluster2 is 5.0; Tried FastSCP, and getting the same result as I see with direct SCP with compression: about 0.5MB/s.
EDIT 2: Increased system resource allocation on VMware hosts to the levels that VMs are getting and beyond. Only change is SCP with compressing from host to host is not up from 0.5M to 4M, just like non-compressed transfer.
Still puzzling why would host-to-host transfer be slower though.
EDIT 3: after adding more resources, was able to achieve 4~10MB/s transfer speeds across data centers. Even though it is lower than 20~40 that the network should be capable of, I'll just have to live with that. Although if anyone has any other ideas - I'm eager to try them :)
With new memory entitlement in vSphere 5, if I have 2 clusters configured with the same license level. Would my memory entitlement be calculated separately for each cluster, or will it pool everything together (managed by the same vCenter server)
I plan on having one Enterprise host, and a few Standard licensed hosts. Will I be able to vmotion between these 2 types of servers? I know that the memory pool will be counted separately, but can't find any information on vmotion restrictions.
I'm trying to figure out how EQL works with multiple controllers. Dell sales people recommended having one RAID array on one EQL unit. But will this set up allow utilizing both controllers? Or will the throughput be limited to 2G, with IOPS limit of a single controller in that setup?
I know that for MD3000i we were advised to set up 2 RAID arrays, so that primary path can be set in a way that will utilize both controllers, and all 4 of the NICs.
So how's it done with EQL?
WARNING: SSLv3 is obsolete. Consider disabling it altogether.
I'm trying to set up Stunnel to server as SSL cache. Everything was smooth, and mostly it works as designed.
Then I encountered errors in log files:
SSL_accept: 1408F10B: error:1408F10B:SSL routines:SSL3_GET_RECORD:wrong version number
Not all clients trigger that, for some strange reason. Connecting from CentOS using links - error shows up (tried multiple machines). Connecting from Ubuntu using links - no error.
Tried using wget, and all is smooth with TLSv1, but error shows up with SSLv3. At the same time, wget reports:
OpenSSL: error:14094410:SSL routines:SSL3_READ_BYTES:sslv3 alert handshake failure
Unable to establish SSL connection.
Here is my config:
pid = /etc/stunnel/stunnel.pid
debug = 3
output = /etc/stunnel/stunnel.log
socket=l:TCP_NODELAY=1
socket=r:TCP_NODELAY=1
verify=3
; fixing "fingerprint does not match" error
fips=no
[https]
accept=12.34.56.78:443
connect=127.0.0.1:80
TIMEOUTclose=0
xforwardedfor=yes
CAfile = /path/to/ssl/example.com.cabundle
cert=/path/to/ssl/example.com.crt
key=/path/to/ssl/example.com.key
Anyone has any insight as to what's happening here? Googled for multiple hours now, can't figure it out.
OpenSSL version: OpenSSL 0.9.8e-fips-rhel5 01 Jul 2008.
Stunnel version: 4.32
EDIT:
Here is an output of openssl s_client -connect example.com:443 -ssl3
CONNECTED(00000004)
3897:error:14094410:SSL routines:SSL3_READ_BYTES:sslv3 alert handshake failure:s3_pkt.c:1086:SSL alert number 40
3897:error:1409E0E5:SSL routines:SSL3_WRITE_BYTES:ssl handshake failure:s3_pkt.c:530:
Same with -tls1 works OK, without any errors.
I'm trying to set up VMware HA for the cluster, and having trouble understanding how resource monitoring works. We overcommit memory as a general practice, so out provisioned memory is always ~1.5 higher on individual VMs.
So I created a cluster with 2 Hosts in it, and one was ~90% full in terms of memory (and I mean used memory, as provisioned was ~140%). Second host ran no VMs. Tried powering on one VM - and I got an error, saying that would make it impossible to tolerate one host failure.
Reading more, I found that when this happens and you disable the policy to prevent power on, VMware will not guarantee failover for all hosts.
I'm looking into ways to balance the load on our MySQL infrastructure, and can't seem to find an answer that works for me... :)
So, I have one big and fat server, that handles everything. Many DBs, many reads, many writes, etc. It handles it pretty well, but it is a single point of failure.
We've set up a couple of slaves to redirect reads to them, but faced 2 problems: it takes a lot of effort to rebuild all programs to split reads and writes; and sometimes slaves get behind, which leads to very interesting artifacts in the application.
Problems with slaves getting behind: because many databases are mixed - there are both heavy 10-20 minute queries done on data mining side, as well as atomic queries that take no time. But Slave runs one query at a time, so all atomic queries have to wait until heavy one finishes.
To resolve these 2 problems, I was thinking about something like a proxy, that would consider this:
One problem that still remains, but which I want to consider - is fail-over. If master fails - would be nice if slave would take responsibilities of a master, and when master is back up - it would become a slave.
Any pointers to RTFM or case studies on this subject would be welcome =)
EDIT: Googled some more, and in addition to Tungsten enterprise - found dbShards, and Schooner. While looking deeper into that - does anyone have experience with these solutions? Any feedback?
I'm in weird situation, where I bought 3TB drives, only to discover that server's bios does not support it (should have done my homework)...
But I still need to install them in the server, and boot from them. I was thinking there should be some way to format the disk or something like that so that BIOS would work with it. But that does not seem to be a popular topic, so search does not yield many info...
The set up is like this: I have 3 HDD slots, and I intend to install CentOS onto the drives in a software raid 5 configuration. Any idea how I can do it? When partitioning at install time - I'm given a message that bios won't load (even if splitting into under 2TB partitions), and can't proceed.